SIMS2015 Session FN1-MoM: Fundamentals

We report here, the first experiments on ion emission from supramolecular ensembles deposited on free standing graphene. The analysis of isolated small nano-objects (few nm) and supramolecular ensembles by SIMS has two important challenges: a) extremely small amount of analyte, and b) influence of the substrate on the emission. We have used a new set-up to overcome those problems. The set-up comprises the measurement of secondary ions in the transmission direction via the event-by-event bombardment-detection mode, where free standing graphene provides a quasi-immaterial support for the analyte (Suppl. Fig.1). The projectiles used were 25 and 50 keV C₆₀ ions and 520 keV Au₄₀₀⁴⁺. The event-by-event bombardment-detection mode allows to select specific impacts on free-standing graphene from those on the molecular ensembles, at the exclusion of signals from the target holder and support. We show below that the interaction of cluster projectile with molecules deposited on graphene involves new sputtering-ionization pathways.

One set of experiments dealt with C₆₀ molecules deposited as a ~1 molecular layer on 2 layer graphene. The mass spectrum obtained with impacts of 50 keV C₆₀²⁺ (Fig.2 Suppl.) contains peaks of small carbon negative cluster ions and intact molecular ion of C₆₀^- with yield of 1.7%. This yield is ~85 times higher than the yield measured for the emission in the reflection direction from a bulk sample of C₆₀. We have examined different types of molecules (e.g. Phenylalanine), deposited on graphene as: a) isolated molecules, b) nano-size agglomerates of molecules, and c) thin (~nm) uniform molecular layers (Fig.3,4). The yields of molecular ions are high for all types of deposition [10% for (c), 2% for (b), and 0.2% for (a)]. The emitted molecular ions are deprotonated for the cases (b) and (c), but for (a) the intact molecular ions have been detected, which indicates differences in the mechanisms of sputtering-ionization.

This presentation will focus on the mechanism/s of sputtering-ionization in the projectile-graphene-analyte complex with emphasis on the time evolution of the rim around the impacting point (rim temperature and graphene movement).

Work supported by NSF grant CHE-1308312.

The enhancement of sputtered ion yields by oxygen is both a boon and a curse of SIMS. It is a boon because the effect is responsible for the impressive sensitivity and low detection limits of SIMS analysis. But oxygen can also be a curse in that varying oxygen levels, arising in particular from changes in the target sputter yield [1], give rise to matrix effects that can produce yield changes of several orders of magnitude for a given species in different materials, a headache for quantitative analysis in complex samples. One unanswered question is whether the oxygen level alone determines the matrix effect as suggested in [1], or to what extent the chemistry of the substrate contributes. We have evaluated useful ion yields, UY, (ions detected/atom sputtered) for three targets, X (= Si, Al and Cu) and for implanted species in these targets, using ¹⁸O implants as internal standards to determine the O/X ratio at the surface of the sputtered material during sputtering. This allows us to compare ionization probabilities for various species in different matrices at identical oxygen concentrations. The results demonstrate that the nature of the substrate can play a significant role in ionization – for example the UY for Cu⁺ sputtered from an oxidized Al surface is an order of magnitude higher than for Cu⁺ sputtered from oxidized Cu at the same oxygen concentration. Intriguingly, also, certain chemically reactive species in less-reactive substrates appear to “getter” low levels of oxygen, so that their ion yields can be almost independent of the nominal oxygen content of the substrate.

[1] "Mechanism of the SIMS matrix effect", Vaughn Deline, William Katz, Charles A. Evans, Jr., Peter Williams, Appl. Phys/ Lett. 33 (1978) 832-835

The effect of the element oxygen on the sputtering process has been examined in molecular dynamics (MD) simulations of C₆₀ bombardment of the polymer polylactic acid (PLA). The atomistic ReaxFF potential [1] has been used in a parallel version of the LAMMPS [2,3], MD code modified for sputtering simulations. The sample consisted of 1 062 891 atoms and the simulation was run for 50 ps. A plethora of small stable molecules are observed both to eject as well as be deposited in the sample. Some of the molecules such as CO and CO₂ appear obvious from the structure of PLA. These molecules are formed from the combinations of adjacent atoms as well as combinations of atoms that originated from different points in the substrate. Other molecules formed are not obvious from the PLA structure such as water, H₂O, and 1-propen-1-one (C₃H₄O). The results will be compared to our simulations of C₆₀ bombardment of the polymer polystyrene (PS) using the same atomistic potential as well as short-time simulations of C₆₀ bombardment of polystyrene using AIREBO potential from the Delcorte group. [4] In addition, the results will be compared to experimental results from the Penn State group. [5]

[1] L. Liu, Y. Liu, S. V. Zybin, H. Sun and W. A. Goddard, Journal of Physical Chemistry A 115 (2011) 11016.

[2] S. Plimpton, J Comp Phys, 117 (1995) 1.

[3] H. M. Aktulga, J. C. Fogarty, S. A. Pandit, A. Y. Grama, Parallel Computing 38 (2012) 245.

[4] A. Delcorte, V. Cristaudo, V. Lebec, B. Czerwinski, International Journal of Mass Spectrometry 370 (2014) 29.

[5] L. Breuer, N. Popzcun and N. Winograd, This Conference.

This talk presents recent theoretical advances aimed at understanding processes taking place during cluster bombardment of organic materials. In the last few years several accomplishments have been achieved. A quest for finding appropriate set of reduced variables has brought us closer towards establishing the basis of the so called “universal representation” of sputtering yield [1 , 2]. Development of a hybrid approach where results of short time computer calculations are fed as start parameters into analytical models has enabled to overcome time limitations of molecular dynamics simulations. As a result, phenomena taking place during depth profiling of organic samples could be studied [3]. Finally, implementation of parallel computational schemes combined with new interatomic potentials makes it possible to perform atomistic studies on a wide variety of organic systems, starting with samples composed from hydrocarbon molecules like β-carotene [4], through peptide systems [5], up to materials composed from long polymer systems [6, 7]. Recent results of such simulations are discussed. The effect of the projectile and material properties on the sputtering of organic materials by cluster projectiles is elucidated in order to identify the main processes responsible for molecular emission and sample modification. The emphasis is placed on phenomena important to the SIMS/SNMS community.

[1] M. P. Seah, J. Phys. Chem. C 117 (2013) 12622.

[2] R. J. Paruch, B. J. Garrison, M. Mlynek, Z. Postawa, J. Phys. Chem. Lett. 5 (2014) 3227; R. J. Paruch, Z. Postawa, B. J. Garrison, this conference.

[3] R. Paruch, B.J. Garrison, Z. Postawa, Analytical Chemistry 85 (2013) 11628.

[4] G. Palka, M. Kanski, D. Maciazek, B. J. Garrison and Z. Postawa, Nuclear Instrum. Meth. B 352 (2015) 202.

[5] C. Mucksch,, C. Anders, H. Gnaser, H.M. Urbassek, Journal of Physical Chemistry C 118 (2014) 7962.

[6] A. Delcorte, V. Cristaudo, V. Lebec, B. Czerwinski, International Journal of Mass Spectrometry 370 (2014) 29

[7] M. Kanski, D. Maciazek, B.J. Garrison, Z. Postawa, this conference; M. Kanski, Z. Postawa, B.J. Garrison, this conference.

The universal descriptions of physical processes are commonly used for simple systems like noble gases and liquids that consist of spherical molecules. Recently the same language has been applied to the dependence of the sputtering yields due to energetic cluster bombardment on the energy of the incident cluster. The assignment of universality has been based on the observation that the spread in data points is reduced when the yield Y and projectile cluster kinetic energy E are expressed in quantities scaled by the number of the cluster atoms n, which is Y/n versus E/n.^[1,2] The alignment of the data points is, however, not perfect and there are two distinct arms of this dependence for organic (molecular) and inorganic (atomic) solids.^[2] The physics underlying the universal dependence is not fully understood, nor is the basis for scaling the variables.

We have demonstrated that the alignment between the data points for organic and inorganic solids can be improved even further if the solid binding energy U₀ is included and the dependence is presented in the form of Y/(E/U₀) versus (E/U₀)/n.^[3] The objective of this study is to address the physical basis of this dependence and determine if such a representation can be universal among the organic and inorganic solids. The ultimate question is how the result of this analysis can be applied to optimize the experimental conditions of the secondary-ion mass spectrometry (SIMS) measurement. For this study we employed the molecular dynamics simulation of Ar_n (n = 60÷2953) cluster bombardment of organic (benzene, octane and β-carotene) and inorganic (Ag) solids for the experimentally accessible range of the cluster beam kinetic energy, at normal incidence.

Our results show that for almost three orders of magnitude variation of (E/U₀)/n, there are obvious similarities in the ejection mechanisms between the organic and inorganic solids, which supports the statement of universality. The representation that includes only U₀ cannot, however, be universal because U₀ cannot alone represent all the material and cluster properties that influence the sputtering phenomena. Moreover, neither the Y/n or Y/(E/U₀) representation includes the energy loss physics associated with fragmentation of molecules. The talk will describe the mechanisms associated with the different ranges of (E/U₀)/n and how understanding the mechanisms aids in understanding the optimal conditions for SIMS experiments.

[1] C. Anders, H. M. Urbassek, R. E. Johnson, Phys. Rev. B 2004, 70, 155404.

[2] M. P. Seah, J. Phys. Chem. C 2013, 117, 12622.

[3] R. J. Paruch, B. J. Garrison, M. Mlynek, Z. Postawa, J. Phys. Chem. Lett. 2014, 5, 3227.