Angle Dependence of Argon Gas Cluster Sputtering Yields for

Mariavitalia Tiddia, Ichiro Mihara, Martin P. Seah, Gustavo Ferraz Trindade, Felix Kollmer, Clive J. Roberts, Richard Hague, Guido Mula, Ian S. Gilmor...
0 downloads 0 Views
Download PDF
Article pubs.acs.org/JPCB

Angle Dependence of Argon Gas Cluster Sputtering Yields for Organic Materials M. P. Seah,* S. J. Spencer, and A. G. Shard Analytical Science Division, National Physical Laboratory, Teddington, Middlesex TW11 0LW, United Kingdom ABSTRACT: The first angle-dependent measurements of the sputtering yield of an organic material using argon gas cluster ions under a wide range of conditions are reported in order to develop an analytical description of the behavior important for the development of the application of secondary ion mass spectrometry to organic and biological systems. Data are presented for Irganox 1010 using argon gas cluster ion beams of 5 and 10 keV energy, E, with cluster sizes, n, from 1000 to 5000. The measurements are conducted in an X-ray photoelectron spectrometer for a range of angles from 0 to 80° from the surface normal. The results support the Universal Equation for argon gas cluster sputtering yields with the angle dependence incorporated into the equation via a simple angle dependence of the parameter A. This explains how and why the angular dependence of the sputtering yield changes significantly with increasing E/n. These results are also accurately confirmed using the published measurements for polystyrene by Rading et al.

1. INTRODUCTION The sputtering behavior of organic materials is important for the analysis of modern electronic devices as well as for the location of drugs and other chemicals in biological systems studied by secondary ion mass spectrometry (SIMS). A breakthrough came in 20101 with the demonstration of the use of argon gas cluster ion sources as the sputtering ion in SIMS. The infrastructure for understanding and describing this sputtering behavior is still in its early days. The present study is part of that infrastructure. The sputtering yields of solid materials, studied for their compositional depth profiling by Auger electron spectroscopy (AES), X-ray photoelectron spectroscopy (XPS), and SIMS, have been a major focus for many years.2−5 The yields for elemental materials using monatomic inert gas primary ions vary strongly with the material sputtered, but semiempirical equations (somewhat complicated) are available that predict sputtering yields to an accuracy of some 10%.6,7 Sputtering yield predictions for simple compounds are more complex8 and less accurate. The angle dependence of the sputtering yield for monatomic ions was observed by Wehner9 in very early studies. Yamamura et al.10 reviewed earlier work and developed a detailed set of equations so that the effect of the incident angle on the sputtering yield could be calculated. More recently, Seah11 showed how these angular variations control the development of topography and the evolution of sputter depth profiles for large nanoparticles. The basic dependencies for the sputtering yields that occur for elemental samples bombarded by monatomic inert gas ions have extensive literature and are thus relatively well-understood in terms of the basic physical processes for predicting practical behavior. In contrast, the equivalent understanding and description for organic materials is just beginning. In recent years, through the evolution of a number of cluster ion beams,12 argon gas clusters have been developed for depth Published XXXX by the American Chemical Society

profiling organic materials. These show either no molecular damage (fragmentation or cross-linking) or minimal damage in the profile using SIMS13 or XPS,14 and excellent depth resolution over significant depths is achieved in both polymeric and organic materials.1,13,15 Owing to the many-body aspects of the collision process, the types of approximations available in the binary collision model used to evaluate monatomic sputtering yields are not available. Instead, molecular dynamics calculations have been made for specific conditions with considerable success,16−18 but it is very difficult to extend any results from one condition to another. Until recently, there was significant discrepancy between the sputtering yields estimated by molecular dynamics simulations and experimental results, particularly for cluster primary ions with a large number of atoms. Largely, this was because the simulations were generally conducted with the primary ion impacts at normal incidence while experiments were typically performed with oblique (∼45°) impacts. The discrepancy is being addressed, but these simulations as yet provide no clear description of sputtering yields as a function of impact angle, cluster size, and energy as part of a dynamic profile. In this paper, we demonstrate that a unified description of sputtering yields of the model compound Irganox 1010, using argon cluster primary ions, can be found in a consistent and simple manner and that the sputtering yields of polystyrene can be similarly described. The implication of this finding is that the sputtering yield of organic materials under any given condition may be predicted using only five material-dependent parameters. This permits the rapid development of experimental and theoretical understanding of cluster ion sputtering through identifying the minimum number of parameters required to fully describe any Received: December 12, 2014 Revised: January 16, 2015

A

DOI: 10.1021/jp512379k J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B

Article

Angular measurements for the smaller coronene cluster primary ions have been made for Irganox 1010 by Seah et al.24 for 8 and 16 keV beam energies between 0 and 80° to the surface normal. Those measurements were made by XPS using reference films of known thicknesses with Irganox 3114 marker layers by monitoring the marker layer signal. It was found that the grazing incidence beam did not deposit significant damage in the Irganox 1010 and that the angular dependence was weak up to ∼75° and not very dependent on the beam energy. Beyond 70°, the yield rapidly fell, extrapolating to zero at 90°. Angular dependence measurements are easier to make in XPS instruments because, unlike SIMS, there is no extraction field around the sample, and therefore, sample tilting does not affect the signal detection. For this reason, XPS is used in the present work.

particular system. It also permits predictions to be made for depth resolutions and spectral changes with the incident beam angle. In a very early study of seven elemental materials sputtered using argon cluster ion beams, Yamada et al.19 show that the sputtering yield is approximately inversely proportional to the binding energies of the atoms. This is similar to the case established for monatomic sputtering. In 2013, Seah20 proposed a generic equation for sputtering yields, Y, for argon gas clusters of energy E with n atoms in the cluster B(E /An)q Y = n 1 + (E /An)q − 1

(1)

where the coefficient B is unity when the yield is measured in atoms per primary ion and is on the order of 0.01 nm3 when the yield is measured as a volume in nm3. It has been found helpful for organic materials, in practice, to measure the yield as a volume. Equation 1 was found to describe the yields of Au, Si, SiO2, Irganox 1010 (C73H108O12 with molecular mass 1176.78 Da), the model organic light-emitting diode material HTM-1 (2,2′,7,7′-tetra(N,N-ditolyl)amino-9,9-spiro-bifluorene, C81H68N4 with molecular mass 1096.54 Da), polystyrene, polycarbonate, and poly(methyl methacrylate), as measured in a range of laboratories. These involved energies from 2.5 to 40 keV and cluster sizes from 100 to 10000 atoms. The important parameter, A, was found to be high for the elements and inorganic compounds but low for organic and polymeric materials. It was found that A was related to U, the energy per atom, excluding hydrogen, to remove the fragment from the solid. For elements, U is on the order of 3−4 eV, but for organic materials with 10−20 carbon atoms in typical fragments, this energy is 10−20 times smaller, and therefore, the value of A was that much smaller. The average A value for the elements and inorganic materials was ∼50 eV, and for the polymers and inorganic materials, it was ∼3 eV. The parameter A is thus an analogue to the binding energy, U, used for calculating monatomic ion sputtering yields. The parameter q is also important and was typically ∼3. For E/n values less than A, eq 1 shows how the yield starts to fall more rapidly with the power of dependence being (E/n)q. In practical situations, elemental and inorganic solids are generally in this E/n < A regime, whereas polymers and organics, with their lower A values, are more in the linear regime where Y is approximately proportional to E.20 A further study of the parameter A was made by Seah21 based on the data of Cristaudo et al.22 for the sputtering yields of polystyrene and poly(methyl methacrylate) as their molecular weights were reduced. It was found that the data were welldescribed by eq 1 with A reducing with the molecular weight according to the average linear size of the molecule in three dimensions; A fell as the energy per atom to remove the fragment from the solid fell, as required in the formulation of eq 1. Practically all of the above sputtering yield measurements are made for argon cluster ions at 45° to the surface normal, and therefore, eq 1 has only been tested at that angle. The only detailed angular measurements for argon clusters have been made for polystyrene by Rading et al.23 for Ar2000 at 5 and 10 keV beam energies at angles between 0 and 75° to the surface normal. Those measurements were made by profilometry of craters in thick films sputtered in a SIMS instrument. For both of the energies, the yield maximized at around 45°.

2. EXPERIMENTAL SECTION Depth profile measurements were conducted in a Kratos Axis Ultra DLD X-ray photoelectron spectrometer to which an Ionoptika argon cluster ion gun, GCIB 10S, had been added at 65° to the spectrometer input lens axis. The gun contains a Wien mass filter that was set to select a narrow range of argon cluster sizes. The X-ray monochromator was set at 60° to the input lens axis in the azimuth opposite to that of the ion gun. The ion beam incidence angle was set by tilting the sample on an axis that was normal to the azimuths containing the X-ray monochromator and the ion gun. In this instrument, the sample heating and cooling system is of an early design in which the sample tilt, unfortunately, does not match the indicated tilt setting as a result of using a magnetically coupled rotary drive to set the tilt angle. The tilt setting was recalibrated, as discussed earlier.24 Many samples of 10 mm squares of silicon wafer, with approximately 50 nm of Irganox 1010, were prepared as described earlier.25 Two extra samples of approximately 100 nm were also made. The thicknesses were measured by spectroscopic ellipsometry using a Woollam M-2000DI instrument. Profiling for many of the samples was conducted by sputtering the sample for a short period at the required angle of incidence of the argon beam, following which the sample was reset with its surface perpendicular to the spectrometer input lens axis, for the reason noted below, and the XPS measurement made. For the sputtering angles of incidence of 45, 60, and some of the 70° measurements, the sample was left tilted at 20, 5, or −5° (the positive angle is toward the ion gun azimuth), respectively, during the XPS measurement to save time in the profiling. However, for sputtering at angles of incidence less than 35°, the samples need to be reset, or the Xrays would strike the back of the sample. A typical sputtering dose between XPS measurements was ∼0.05 ions/nm2. This process was repeated until the Si substrate was observed. The XPS source was of monochromated Al Kα X-rays that illuminate approximately an area 1 mm × 2 mm on a sample when set with its surface perpendicular to the spectrometer input lens axis. Within this area, the analyzer was set to analyze a region of only ∼220 μm diameter to obtain good depth resolution. The carbon 1s and silicon 2p regions were measured using the instrument in the snapshot mode, where ∼0.1 eV steps for ∼14.5 eV of the binding energy scale were recorded simultaneously. The charge neutralizer and X-ray source were only used during the acquisition of spectra, both being turned off during the sputtering cycle. The ion beam was rastered to sputter a region on the sample with sides from 1.4 to 2.6 mm B

DOI: 10.1021/jp512379k J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B

Article

by 3.6 mm for angles of incidence up to 60°. Above this angle, the raster size became an elongated rectangle up to 6.4 mm × 3.6 mm. The sputtered areas were measured optically from each crater. The ion beam current was measured before and after each depth profile using a Faraday cup26 mounted parallel to the direction of the ion beam. The average change in these two measurements of the beam current was 10%. The sputtering dose rate was calculated directly from the average measured beam current and the measured crater dimensions for each angle.

3. RESULTS AND DISCUSSION 3.1. Irganox 1010. In conducting the depth profiles, the C 1s and Si 2p signals, shown in Figure 1, were used. Note the Figure 2. Depth profile showing the C 1s and Si 2p intensities as a function of Ar gas cluster sputtering dose for a 49 nm thick Irganox 1010 layer sputtered with 10 keV Ar5000+ ions incident at 15° from the surface normal.

dose to the centroid of the interface (50% level) to be established very precisely. The linear fall before the interface is associated with the current state of the X-ray monochromator and was not associated with sample degradation. Depth profiles with argon gas cluster ion beams in SIMS show no significant changes in the profiles. In the measurement of profiles using XPS, each atom gives a signal whose intensity can be seen through the overlying material to an extent governed by the attenuation length, L, which, for the Al Kα X-rays used here and for typical organic materials, is 3.0 nm.24,31,32 With no broadening of the interface, the C 1s intensity will fall as I0{1 − exp[−x/(L cos α)]}, where I0 is the intensity away from the interface, x is the distance of the surface from the interface, and α is the angle of emission of the photoelectrons detected by the analyzer. This effect shifts the apparent centroid observed by a distance L cos α toward the surface.24,28 In the calculations that follow, this quantity is subtracted from the measured film thicknesses to calculate the sputtering yields. Thus, the yield volume for the 10 keV Ar5000 data, shown in Figure 2, where the film thickness, d, is 49 nm and the dose, D, to the interface is 2.5 ions/nm2, is given by (d − L cos α)/D, which is 18.4 nm3. The measured yields are thus shown in Figure 3a. There is some scatter in the data, but there are also clear trends. The data from the 100 nm films showed results not statistically different from those for the 50 nm films, indicating the linearity of the sputtering with dose. This linearity is already wellestablished in SIMS13,30 in profiling samples with delta layers at known depths. The yield is highest at around 45° incidence, and the enhancement over the yield at 0° can range from 2 using 10 keV Ar1000+ (top trace) to 10 using 5 keV Ar2000+ (bottom trace). It is clear that the Universal Equation shown in eq 1 is not to be modified by simply replacing the factor B by a simple angle-dependent factor, B(θ). The situation is a little more complex than that. Figure 3b homogenizes the data somewhat by grouping all of the data by the E/n values. We noted earlier20 that the Universal Equation, eq 1, indicates that if we double the value of n and double E, then the yield simply doubles, that is, there is linearity in the yield if we add more atoms of the same velocity to the primary ion cluster. If the behavior is so governed by E/n, it is reasonable to expect that the angle dependence is also governed by E/n. This is what Figure 3b shows; as E/n falls, the angle dependence gets

Figure 1. X-ray photoelectron spectrum from midway through a profile using 10 keV Ar1000+ cluster ions. The Ar 2p signal is within the noise level.

lack of observable implanted argon. Here, it is less than 0.5%.5,27 In the studies of elements using monatomic argon,6 typically 1−5% was implanted, but that did not occur here. The intensities of these peaks were measured from the integrated peak areas above a straight line background from the narrowscan, snapshot mode, data. In this instrument, profiles were made by tilting the sample to the angle for sputtering, and ∼1 nm of material was removed using a small sputtering dose. The sputtering was then stopped, and in many cases, the sample was tilted to the horizontal position for XPS measurements. This cycle was repeated some 50 times. The depth profile traces, exampled in Figure 2, were not, therefore, continuous and required many sputtering and XPS cycles with the sample tilt reset each time. The interfaces are so sharp that detail of the interface is, unfortunately, lost, but the interface dose is very well-defined. During sputtering, the Irganox 1010 showed no chemical degradation with the spectra closely resembling Figure 5a, b, and d of Seah et al.24 Had there been significant detail at the interface, it would probably have been well-described by the integral of the asymmetric function described by Dowsett28,29 and by Shard et al.,25 which has been found to describe the broadening of delta layers and roughening in the sputtering of these organic materials.13,25,30 Here, however, it is adequate to fit this interface by the integral of a Gaussian resolution profile, as shown by the solid black line with the C 1s data in Figure 2. Note that the absolute intensity of the C 1s peak falls roughly linearly with the dose, and therefore, a linear fall is incorporated in the fit including the integrated Gaussian. This allows the C

DOI: 10.1021/jp512379k J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B

Article

Figure 4. Values of 1/[A(θ)] eV−1 deduced for each angle, applying eq 1 with q = 3.3 and B = 0.0176 nm3. The solid curve is for eq 2 with A(0°) = 4.1 eV, F = 3.1, and θo = 18°.

4. However, for E/n < A(θ), Y(θ) ≃ nB{E/[nA(θ)]}q, that is, Y(θ) ≈ [1/A(θ)](q). We noted earlier that q = 3.3, so that Y(θ) ≈ [1/A(θ)]3.3 and the angle dependence, as E/n falls below ∼3, becomes much stronger than the curve of Figure 4. The solid lines in Figure 3b show this effect using eq 2. We may view eq 2 as arising from emission with two roots, (i) from bond breaking and ejection from the component of the energy in the impact normal to the surface and (ii) from a lateral, more forward sweeping action parallel to the surface that increases as θ increases but also then declines as reflection of the primary ion from the surface increases at high θ. This simple model would combine with E (or equivalently 1/A) to generate angular dependencies approximating p cos2 (θ) and q cos2 (θ) sin2 (θ), respectively, with the coefficient p weaker than q if the sweeping action is the stronger. For p = 0.1q, the curve describes the essential behavior of the data but is less accurate than eq 2; therefore, in all of the discussions, eq 2 is used. The fits of the full data set using eqs 1 and 2, plotted to illustrate the Universal Equation, are shown in Figure 5a. In this format, for any given angle, the data for many combinations of E and n fall on one curve. The data show a relative scatter of 17% about the curves that reflects the data repeatability. It is clear that the sputtering yields for 45° are the highest for all angles at high E/n, and the variation with angle becomes relatively stronger as E/n falls, as described above. Figure 5b shows how, using an abscissa of E/[nA(θ)], the Irganox 1010 sputtering yield data at all angles now fit one curve for the Universal Equation. Earlier, we published data for Irganox 1010 measured at 45° by Niehuis et al.15 and by Shard et al.30 Figure 5c reproduces those data and shows the fit of the current equation with slightly different q and A(45°) values from those given previously.20 It is from the fitting to all these data that the value for q of 3.3, used earlier, was obtained. It is clear that the Universal Equation combined with the angular dependence described by eq 2 is good for all of the Irganox 1010 data. Equation 2 conveniently separates the angular and energy terms

Figure 3. Angle dependence of the sputtering yield of Irganox 1010 for 5 of the E and n combinations, (a) the yield data, (b) the data normalized by Y(0°) calculated theoretically for each E/n value and with the data grouped by E/n value. In (b), the solid lines are the predictions using eqs 1 and 2, as described later.

stronger and stronger. To bring the curves together in Figure 3b for presentational purposes, the data have been divided by a calculated Y(0°) value for each E/n, which will be described later. The simplest way of including an angle dependence in eq 1, which has the behavior required, is through the energy parameter A. We now call this A(θ). We may use eq 1 with B = 0.0176 nm3 as found previously20 for Irganox 1010 and evaluate A(θ) for each angle. We shall put q = 3.3 here for reasons discussed later. Figure 4 shows the results for 1/[A(θ)] eV−1 as a function of angle from these data. The results fit a simple curve showing that A(θ) is smallest near 45° (1/[A(θ)] is highest), leading to high yields and then lower yields at both higher and lower angles. Clearly, the popular experimental arrangement of θ = 45° gives the most efficient sputtering. Superimposed in Figure 4 is an empirical analytical expression to describe the yield ⎡ 1 + F exp( −θ 2/2θ 2) ⎤ o ⎥ A(θ ) = A(0°)⎢ ⎣ (1 + F ) cos(θ ) ⎦

(2)

Equation 2 describes the measured A(θ) values to better than 3% with only three coefficients. Here, A(0°) = 4.1 eV, F = 3.1, and θo = 18°. The way that A changes the yield behavior of eq 1 is illustrated in some detail by Seah et al.33 Briefly, for E/n > A(θ), eq 1 shows that Y(θ) ≃ BE/A(θ) so that the angular dependence is reflected directly by the curve shown in Figure

A(θ ) = A(0°)C(θ )

(3)

where C(θ) contains the angular terms in the square brackets in eq 2. One result of the change in A(θ) with angle that may not be expected is that the relative intensities of secondary ions in D

DOI: 10.1021/jp512379k J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B

Article

equilibrium.36,37 Nevertheless, we can see in the molecular dynamics results that as E/n falls below around 5 eV, the sputtering peak is around 45° and that the yield at small angles drops dramatically34,35 in a similar manner to that shown in Figure 3. Rzeznik et al.34 note that it is surprising that the impact angle dependence is quite different in the case of large projectiles (n = 2953 compared with n = 60 at 15 keV), which cannot be explained by deposition of the primary energy below the critical depth. What we can now see is that the change increases smoothly as the E/n value falls. When E/n is high, the angular variation below 60° is weak, and Y(45°)/Y(0°) ≈ A(0°)/A(45°) is typically 2, but once E/n falls below A(0°) (=3.9), this ratio rises and, below A(45°) (=1.6), it rises typically to 10. In this lower E/n regime, the yield is in the power law regime. As the energy falls, as noted in the molecular dynamics simulations of Delcorte and co-workers,16,38−40 it is shown that below 1 eV/nucleon (i.e., 40 eV/Ar atom), the sputtering process changes from being more atomistic with bond breaking to a situation in which the primary ion cluster remains essentially intact and the fragmentation induced in the molecule is reduced. This change is actually observed at around 4 eV/Ar atom, experimentally, in several organic molecules, and although there is a difference of almost a factor of 10 between the molecular dynamics simulations and experimental data, the explanation still seems valid. This is the energy at which the sputtering switches from the linear to the power law regime. Equation 2 is very important for practical reasons because it allows the calculation of the yield for many conditions. We know that the SIMS ion intensities and the depth resolution both change with the values of E and n, but now, we have an added degree of freedom to improve these for depth profiling or imaging because, as we see from Figure 3b, an order of magnitude reduction in yield is now available for the primary ion beam operated at the same energy with the same cluster size. 3.2. Polystyrene. As mentioned in the Introduction, the only other detailed data on the angle dependence of the sputtering yield using argon gas clusters is that for polystyrene by Rading et al.23 for 5 and 10 keV Ar2000+. Those data, shown in Figure 6a, also give a stronger effect for the lower E/n values. We have previously analyzed their data at 45° for the many E/n values used23 and have shown that they are well-described by the Universal Equation, eq 1. Here, we fit all those data and the angle-dependent results with eqs 1 and 2 with q = 3.9, B = 0.011 nm3, A(0°) = 3.363 eV, F = 2.7, and θo = 29.5° to give the solid curves in Figure 6a that fit the data to 7% rms scatter. The fit of eq 1 to the data fitted by Seah20 with the present A(45°) = 2.37 eV from eq 2 is shown in Figure 6b. Here, the relative standard deviation of the scatter is 9%, just the same as that in the previous fitting20 because the parameters have not changed significantly. These results for polystyrene have high internal consistency because all of the data are recorded using one technique (SIMS) in one laboratory.

Figure 5. Fits of eq 1 (a) to the values of Y/n for the angle-dependent sputtering yields measured with A(θ) values given by eq 2 with A(0°) = 4.1 eV, F = 3.1, and θo = 18°, (b) the data and fits unified by changing the abscissa to E/[nA(θ)], and (c) to the values of Y/n by Niehuis et al.15 and Shard et al.30 with A(45°) = 1.607 eV from eq 2.

single-beam SIMS may change, as described by Seah et al.33 At 0°, compared with 45°, because A(0°) > A(45°), the ratio of E/n to the sample internal bonding in eq 1 has fallen. The total yield falls but so too may the ratio of the small and large emitted secondary ion fragments. It remains to be seen if this predicted enhancement of the large ion fragments is observed in single-beam SIMS at 0 or 80° compared with the regular measurements at 45°. Angular-dependent molecular dynamics calculations are available for benzene using argon clusters at 10 and 15 keV.34,35 These are for the initial impact on a previously unsputtered sample and will not be exactly the same as the present results (ignoring the change in sample) because we do know that the sputtering needs to be considered at

4. CONCLUSIONS The angular dependence of the sputtering yields of Irganox 1010 has been measured in an XPS instrument for 5 and 10 keV argon cluster ion beams with cluster sizes in the range of 1000−5000. The results are described by the new generic description of eq 2 with three fitting parameters. This shows that the angular dependence depends on E/n with an enhancement at 45° that changes between 2 and 10 times that at 0°. The geometry of many XPS and SIMS systems with E

DOI: 10.1021/jp512379k J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B

Article

of Business, Innovation and Skills and with funding by the European Metrology Research Programme (EMRP) Projects NEW01-TReND and IND15-SurfChem. The EMRP is jointly funded by the EMRP participating countries within EURAMET and the European Union.

■

(1) Lee, J. L. S.; Ninomiya, S.; Matsuo, J.; Gilmore, I. S.; Seah, M. P.; Shard, A. G. Organic Depth Profiling of a Nanostructured Delta Layer Reference Material Using Large Argon Cluster Ions. Anal. Chem. 2010, 82, 98−105. (2) Sigmund, P. Theory of Sputtering I. Sputtering Yield of Amorphous and Polycrystalline Targets. Phys. Rev. 1969, 184, 383− 416. (3) Sigmund, P. Sputtering by Ion Bombardment: Theoretical Concepts. In Topics in Applied Physics, Vol 47, Sputtering by Particle Bombardment I; Behrisch, R., Ed.; Springer: Berlin, Germany, 1981; Chapter 2, pp 9−71. (4) Hofmann, S. Depth Profiling in AES and XPS Practical Surface Analysis 2nd ed. Vol. 1  Auger and X-ray Photoelectron Spectroscopy; Briggs, D., Seah, M. P., Eds.; Wiley: Chichester, U.K., 1990; Chapter 4, pp 143−199. (5) Wittmaack, K. Basic Aspects of Sputter Depth Profiling Practical Surface Analysis, 2nd ed. Vol. 2  Ion and Neutral Spectroscopy; Briggs, D., Seah, M. P., Eds.; Wiley: Chichester, U.K., 1992; Chapter 3, pp 105−175. (6) Seah, M. P.; Clifford, C. A.; Green, F. M.; Gilmore, I. S. An Accurate Semi-Empirical Equation for Sputtering Yields, I: For Argon Ions. Surf. Interface Anal. 2005, 37, 444−458. (7) Seah, M. P. An Accurate Semi-Empirical Equation for Sputtering Yields, II: For Neon, Argon and Xenon Ions. Nucl. Instrum. Methods Phys. Res., Sect. B 2005, 229, 348−358. (8) Seah, M. P.; Nunney, T. S. Sputtering Yields of Compounds Using Argon. J. Phys. D: Appl. Phys. 2010, 43, 253001. (9) Wehner, G. K. Influence of the Angle of Incidence on Sputtering Yields. J. Appl. Phys. 1959, 30, 1762−1765. (10) Yamamura, Y.; Itakawa, Y.; Itoh, N. Technical Report IPPJ-AM26. Angular Dependence of Sputtering Yields of Monatomic Solids; Institute of Plasma Physics, Nagaoya University: Nagaoya, Japan, 1983. (11) Seah, M. P. Topography Effects and Monatomic Ion Sputtering of Undulating Surfaces, Particles and Large Nanoparticles: Sputtering Yields, Effective Sputter Rates and Topography Evolution. Surf. Interface Anal. 2012, 44, 208−218. (12) Mahoney, C. Cluster Secondary Ion Mass Spectrometry of Polymers and Related Materials. Mass Spec. Rev. 2010, 29, 247−293. (13) Shard, A. G.; Havelund, R.; Seah, M. P.; Spencer, S. J.; Gilmore, I. S.; Winograd, N.; Mayo, D.; Miyayama, T.; Niehuis, E.; Rading.; et al. Argon Cluster Ion Beams for Organic Depth Profiling: Results from a VAMAS Interlaboratory Study. Anal. Chem. 2012, 84, 7865− 7873. (14) Miyayama, T.; Sanada, N.; Bryan, S. R.; Hammond, J. S.; Suzuki, M. Removal of Ar+ Beam-Induced Damaged Layers from Polyimide Surfaces with Argon Gas Cluster Ion Beams. Surf. Interface Anal. 2010, 42, 1453−1457. (15) Niehuis, E.; Möllers, R.; Rading, D.; Cramer, H.-G.; Kersting, R. Analysis of Organic Multilayers and 3D Structures Using Ar Cluster Ions. Surf. Interface Anal. 2013, 45, 158−162. (16) Delcorte, A.; Garrison, B. J.; Hamraoui, K. Dynamics of Molecular Impacts on Soft Materials: From Fullerenes to Organic Nanodrops. Anal. Chem. 2009, 81, 6676−6686. (17) Delcorte, A.; Cristaudo, V.; Lebec, V.; Czerwinski, B. Sputtering of Polymers by keV Clusters: Microscopic Views of the Molecular Dynamics. Int. J. Mass Spectrom. 2014, 370, 29−38. (18) Paruch, R. J.; Garrison, B. J.; Mlynek, M.; Postawa, Z. On Universality in Sputtering Yields Due to Cluster Bombardment. J. Phys. Chem. Lett. 2014, 5, 3227−3230.

Figure 6. Analysis of the polystyrene data of Rading et al.,23 (a) the angle-dependent data for 5 and 10 keV Ar2000+ together with the predictions of eqs 1 and 2 and (b) the data at θ = 45° for many conditions fitted at 45° using eqs 1 and 2, with the curves at 0 and 70° added for illustrative purposes. Here, q = 3.9, B = 0.011 nm3, A(0°) = 3.363 eV, F = 2.7, and θo = 29.5°.

the ion gun at 45° to the sample surface thus provides the most efficient sputtering geometry. However, this may or may not be the best geometry for high-resolution sputter depth profiling or for maximizing the information content in imaging analysis. The results for Irganox 1010 are all described by an angulardependent parameter A in the Universal Equation of eq 1. The essential energy, cluster size, and angular behavior are fully replicated by eqs 1 and 2. This significantly extends the applicability of the Universal Equation. Analysis of published data for polystyrene shows that this too can be similarly described with an overall relative standard deviation of 7% for both the E/n and θ dependencies of the yield. This indicates that eq 2 may be applicable to the sputtering yield for all organic materials.

■

REFERENCES

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone +442089436634. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors would like to thank Steve Smith for the preparation of the samples and Ionoptika for assistance in the use of the GCIB 10S. This work forms part of the 3D NanoSIMS project in the Chemical and Biological programme of the National Measurement System of the U.K. Department F

DOI: 10.1021/jp512379k J. Phys. Chem. B XXXX, XXX, XXX−XXX

The Journal of Physical Chemistry B

Article

Crosslinking under C60 and Arn Cluster Bombardment. Nucl. Instrum. Methods Phys. Res., Sect. B 2013, 303, 22−26.

(19) Yamada, I.; Matsuo, J.; Toyoda, N.; Kirkpatrick, A. Materials Processing by Gas Cluster Ion Beams. Mater. Sci. Eng., R 2001, 34, 231−295. (20) Seah, M. P. Universal Equation for Argon Gas Cluster Sputtering Yields. J. Phys. Chem. C 2013, 117, 12622−12632. (21) Seah, M. P. Argon Cluster Size-Dependence of Sputtering Yields of Polymers: Molecular Weights and the Universal Equation. Surf. Interface Anal. 2015, 47, 169−172. (22) Cristaudo, V.; Poleunis, C.; Czerwinski, B.; Delcorte, A. Ar Cluster Sputtering of Polymers: Effects of Cluster Size and Molecular Weights. Surf. Interface Anal. 2014, 46, 79−82. (23) Rading, D.; Moellers, R.; Cramer, H.-G.; Niehuis, E. Dual Beam Depth Profiling of Polymer Materials: Comparison of C60 and Ar Cluster Ion Beams for Sputtering. Surf. Interface Anal. 2012, 45, 171− 174. (24) Seah, M. P.; Spencer, S. J.; Shard, A. G. Depth Resolution, Angle Dependence and the Sputtering Yield of Irganox 1010 by Coronene Primary Ions. J. Phys. Chem. B 2013, 117, 11885−11892. (25) Shard, A. G.; Green, F. M.; Brewer, P. J.; Seah, M. P.; Gilmore, I. S. Quantitative Depth Profiling of Organic Delta Layers by C60 Ion Sputtering and SIMS. J. Phys. Chem. B 2008, 112, 2596−2605. (26) Gilmore, I. S.; Seah, M. P. Fluence, Flux, Current and Current Density Measurement in Faraday Cups for Surface Analysis. Surf. Interface Anal. 1995, 23, 248−258. (27) Shard, A. G. Detection Limits in XPS for More Than 6000 Binary Systems Using Al and Mg Kα X-rays. Surf. Interface Anal. 2014, 46, 175−185. (28) Dowsett, M. G.; Rowlands, G.; Allen, P. N.; Barlow, R. D. An Analytic Form for the SIMS Response Function Measured from UltraThin Impurity Layers. Surf. Interface Anal. 1994, 21, 310−315. (29) Dowsett, M. G.; Kelly, J. H.; Rowlands, G.; Ormsby, T. J.; Guźman, B.; Augustus, P.; Beanland, R. On Determining Accurate Positions, Separations, and Internal Profiles for Delta Layers. Appl. Surf. Sci. 2003, 203−204, 273−276. (30) Shard, A. G.; Ray, S.; Seah, M. P.; Yang, L. VAMAS Interlaboratory Study on Organic Depth Profiling. Surf. Interface Anal. 2011, 43, 1240−1250. (31) Seah, M. P.; Spencer, S. J. Attenuation Lengths in Organic Materials. Surf. Interface Anal. 2011, 43, 744−751. (32) Seah, M. P. Simple Universal Curve for the Energy Dependent Electron Attenuation Length for All Materials. Surf. Interface Anal. 2012, 44, 1353−1359. (33) Seah, M. P.; Havelund, R.; Gilmore, I. S. Universal Equation for Argon Cluster Size Dependence of Secondary Ion Spectra in SIMS of Organic Materials. J. Phys. Chem. C 2014, 118, 12862−12872. (34) Postawa, Z.; Paruch, R.; Rzeznik, L.; Garrison, B. J. Dynamics of Large Ar Cluster Bombardment of Organic Solids. Surf. Interface Anal. 2013, 45, 35−38. (35) Rzeznik, L.; Paruch, R.; Garrison, B. J.; Postawa, Z. Sputtering of Benzene Sample by Large Ne, Ar and Kr Clusters  Molecular Dynamics Computer Simulations. Acta Phys. Pol., A 2013, 123, 825− 827. (36) Paruch, R.; Postawa, Z.; Wucher, A.; Garrison, B. J. Steady-State Statistical Sputtering Model for Extracting Depth Profiles from Molecular Dynamics Simulations of Dynamic SIMS. J. Phys. Chem. C 2012, 116, 1042−1051. (37) Garrison, B. J.; Paruch, R.; Postawa, Z. Combined Molecular Dynamics and Analytical Model for Repetitive Cluster Bombardment of Solids. Nucl. Instrum. Methods Phys. Res., Sect. B 2013, 303, 196− 199. (38) Czerwinski, B.; Rzeznik, L.; Paruch, R.; Garrison, B.; Postawa, Z. Effect of Impact Angle and Projectile Size on Sputtering Efficiency of Solid Benzene Investigated by Molecular Dynamics Simulations. Nucl. Instrum. Methods Phys. Res., Sect. B 2011, 269, 1578−1581. (39) Delcorte, A.; Restrepo, O. A.; Czerwinski, B.; Garrison, B. J. Surface Sputtering with Nanoclusters: the Relevant Parameters. Surf. Interface Anal. 2013, 45, 9−13. (40) Czerwinski, B.; Postawa, Z.; Garrison, B. J.; Delcorte, A. Molecular Dynamics Study of Polystyrene Bond-Breaking and G

DOI: 10.1021/jp512379k J. Phys. Chem. B XXXX, XXX, XXX−XXX