Embedding-Free Method for Preparation of Cross ... - ACS Publications

Apr 10, 2017 - and Ian S. Gilmore*,†. †. National Physical Laboratory, Hampton Road, Teddington, TW11 0LW, United Kingdom. ‡. Kuraray Co., Ltd.,...
0 downloads 0 Views 3MB Size
Letter pubs.acs.org/ac

Embedding-Free Method for Preparation of Cross-Sections of Organic Materials for Micro Chemical Analysis Using Gas Cluster Ion Beam Sputtering Ichiro Mihara,*,†,‡ Rasmus Havelund,† and Ian S. Gilmore*,† †

National Physical Laboratory, Hampton Road, Teddington, TW11 0LW, United Kingdom Kuraray Co., Ltd., 2045-1, Sakazu, Kurashiki, Okayama, 710-0801, Japan



S Supporting Information *

ABSTRACT: We present a novel in situ mask method for the preparation of cross-sections of organic materials such as polymer multilayer films suitable for chemical imaging of buried interfaces. We demonstrate this method on a model buried interface system consisting of a piece of Scotch tape adhered to a PET substrate and a protective film used in consumer packaging. A high dose of gallium from a focused ion beam (FIB) was used to produce a damaged overlayer on the surface of the organic sample. The damaged layer has a significantly slower sputter rate compared to the native undamaged organic material. Therefore, during gas cluster ion beam (GCIB) depth profiling experiments the damaged layer functions as a mask, protecting the sample beneath and producing a cross-section at the edge of the mask. The FIB itself cannot be used directly to prepare the crosssection since the organic materials are easily damaged. A four step workflow is described including a final cleaning procedure to remove redeposited material from the cross-section. The workflow is completed in a few hours for samples up to 100 μm thickness. The method does not require sample embedding and is suited to automated analysis, which can be important benefits for industrial analysis where a variety of samples are analyzed routinely.

C

3D SIMS imaging of organic materials is now of great analytical importance. However, a drawback of 3D SIMS imaging is that the reconstructed 3D image, from a series of sequential 2D images, makes an assumption that the sputtering yield is the same for all materials and there are no voids. To a first approximation this is reasonable but sputtering yields may vary by a factor of 10 for organic materials4 and voids are not unusual. A further complication is an explicit assumption that the upper surface is flat. If that is not the case then corrections need to be made usually by assuming that an interface in the imaged volume of sample is flat. This simple correction can work well if such an interface exists.5 Tomographic methods using a focused ion beam (FIB) to mill away and expose a vertical cross-section, like a cliff face, have become popular and can be combined with SEM for highresolution images and also with SIMS for chemical imaging.6 This obviates the difficulties of different sputtering yields since the image of the exposed “cliff face” is taken directly. However, focused ion beams use liquid metal ion sources, and

ross-section analysis is routinely used to study defects in buried layers and interfaces, e.g., segregation of additives or particulate inclusions that lead to delamination in multilayer films or reduced product quality.1,2 Cross-sections are usually prepared by mechanical cutting using a microtome. It is useful for a wide range of samples, including biological samples, but some samples have to be embedded in, e.g., resins, OCT, or wax and require cryogenic conditions to create a smooth crosssection. Those processes can be time-consuming and sometimes trial-and-error due to the variety of samples in industrial analysis. Additionally the operator needs to be well experienced to avoid the introduction of contaminants and artifacts such as smearing at the exposed surface. Secondary ion mass spectrometry (SIMS) is a powerful technique for chemical imaging of cross-sections due to its high sensitivity and submicrometer spatial resolution. It provides valuable information which complements other spectroscopic analysis techniques such as X-ray photoelectron spectroscopy and Raman spectroscopy. Modern SIMS instruments are equipped with gas cluster ion beams (GCIBs), which allow this surface sensitive technique to access deep (tens of micrometers) into organic materials3 since the GCIB sputtering is gentle and does not create subsurface molecular damage. This capability has freed SIMS from the constraints of a surface analytical technique to analyzing deep into organic materials. © 2017 American Chemical Society

Received: February 10, 2017 Accepted: April 10, 2017 Published: April 10, 2017 4781

DOI: 10.1021/acs.analchem.7b00511 Anal. Chem. 2017, 89, 4781−4785

Letter

Analytical Chemistry

Figure 1. Workflow for in situ mask preparation and GCIB cross-sectioning method consisting of four steps: (a) step 1, the FIB creates a userdefined damage region, the mask, shown in red (a rectangular region of 200 μm × 40 μm is selected; (b) step 2, the area selected for high energy ArGCIB sputtering, shown as the green square (500 μm × 500 μm); (c) step 3, cleaning with low-energy Ar-GCIB sputtering is required; (d) step 4, the cleaned cross-section is ready for imaging by TOF-SIMS, SEM, or other analytical techniques.

extraction mode. This set of condition enables a spatial resolution of approximately 200−300 nm to be achieved while the mass resolving power is kept at several thousands. More details are given elsewhere.11,12 Additionally, this mode reduces topographic effects allowing ions from the crater wall to be detected. As a practical and illustrative example, we use adhesive tape of the brand Scotch (3M). The thickness of this adhesive tape is approximately 60 μm, consisting of a 40 μm thick carrier polymer and a 20 μm sticky layer. A thin release layer on top of the carrier layer prevents the tape from sticking too strongly to itself in the tape roll. A piece of adhesive tape was cut and placed on a PET film as a substrate. An insulating substrate is preferred to reduce electric field gradients.13 Workflow for in Situ Mask Method and Analysis of Cross-Section. The four step workflow is shown schematically in Figure 1 along with CCD images from an in situ camera in the TOF-SIMS chamber. In step 1, the FIB is used to create the mask, a region of damaged material (in the top surface). In this example, a rectangular mask (200 μm × 40 μm corresponding to 512 pixels by 103 pixels) was selected; however, alternative patterns may be used. The ion beam current was 20 nA and the total ion beam dose was 3.27 × 1018 ions/cm2. Subsequently in step 2, organic material is sputtered away using the GCIB source selecting 20 keV Ar1700 (E/n ∼ 11.8 eV/atom) with a current of 10 nA. The sputtering is stopped once signals from the PET substrate, such as C7H4O+, are detected and the crosssection is exposed. In step 3, sputter deposited materials on the cross-section are removed by low energy Ar-GCIB sputtering, and this is discussed in more detail later. Finally in step 4, chemical imaging of a freshly prepared cross-section using the preferred analytical technique, in this case SIMS. Owing to the 45° angle of incidence a bevel cut is produced magnifying the cross-section by 1.41, which is convenient. If that was not desired, the sample can be rotated so that the GCIB is normal to the sample surface. The cross-section prepared using the in situ mask method exhibits a straight smooth edge parallel to the incident angle of Ar-GCIB (see the Supporting Information). In contrast, when no mask is used the cross-section is curved and significantly rougher. The top is poorly defined owing to the large beam size of the GCIB. For Scotch tape, this workflow takes 120 min including 30 min for the mask process, 40 min for the sputtering process, and 50 min for the cleaning process.

unfortunately, these severely damage any organic material so that all molecular information is lost. GCIBs do not have the same source brightness and so cannot be focused as tightly. Therefore, there is no direct alternative for organics. Instead, Iida et al. used a thin metal film as a mask to define an edge and then sputtered with an Ar-GCIB to make a crater with a sharp wall.7 Since the sputtering yield, i.e., the volume of material removed by each incoming cluster ion, is 2−4 orders of magnitude lower for metals than for polymers8 even a thin metal film will effectively mask the area it covers. They demonstrate that their method permits molecular SIMS imaging of the cross-section, but there are some clear disadvantages; alignment of the mask is manual and could be tricky to precisely locate the cross-section and consequently is not well adapted to automated control. There is also the risk of introducing contaminants and artifacts during sample preparation. In this short communication, we introduce a novel and simple method where we use a FIB to write a mask on the surface. This creates a region of molecularly damaged, carbonized material that has a sputtering yield 2 orders of magnitude lower than most organics9,10 and so performs the same function as the metal mask used by Iida.7 The important differences are that the mask is in intimate contact with the top surface, it is conformal (though in extreme topography one would need to take care of possible shadowing effects) and importantly the mask can be of arbitrary shape and location defined by the computer system and is therefore easily integrated into defect detection or inspection work flows. This method is well suited for organic multilayer films up to approximately 100 μm thickness. For thicker samples, e.g., millimeter thickness mechanical cross-sectioning would be more appropriate. It is not suited to biological applications where cryo-microtome methods are routinely used. In this study we use a TOF-SIMS 5 (ION-TOF GmbH, Germany) equipped with a dual column ion beam mounted at 45° to the sample surface and with both a 20 keV Ar-GCIB source and a 30 keV Ga FIB source (in line). The time-of-flight analyzer is normal to the sample surface. The instrument is also equipped with a Bi−Mn liquid metal ion beam also mounted at 45° to the sample surface and at 90° polar angle from the dual column. For imaging of the cross-section, 30 keV Bi3+2 was used in the high lateral resolution mode, the so-called “Fast Imaging mode”. The TOF analyzer was operated in the delayed 4782

DOI: 10.1021/acs.analchem.7b00511 Anal. Chem. 2017, 89, 4781−4785

Letter

Analytical Chemistry In order to define the mask, a FIB dose of 3.27 × 1018 ions/ cm2 is found to be required for Scotch tape. The details of this are shown in the Supporting Information. It is likely that this dose is typical for many organic materials of this type. A lower dose could suffice for materials that cross-link more readily such as type II polymers.14 In any case, the damage of polymers and organics by atomic ion beams is well understood.15 We must expect that when sputtering away a significant amount of material using the Ar-GCIB that some residual material will be deposited on the cross-section surface (though the issue is presumably reduced owing to the forward direction of sputtered ejecta16). We show this later. Consequently, step 3 in the workflow is an important one. The sample is rotated 180° so that the cross-section now faces the GCIB and its surface is normal to it. Now a gentle “polish” is given by sputtering with 5 keV Ar1700 (E/n ∼ 3 eV/atom), which has approximately 4 times lower sputtering yield volume compared with the earlier sputtering condition. A lower beam current of 2 nA is also used so that the sputtering rate is less than 40 times lower than the initial sputtering. To measure the effect of cleaning, after every 600 s of sputtering, the stage was rotated again to the original position for SIMS imaging with the Bi3+2 ion beam. The effect of the cleaning procedure is shown in Figure 2. The C2H3O+ ion is

after every 600 s cycle of sputter cleaning. It can be seen that after the first 600 s cycle the deposited C4H9+ (from the sticky layer) on the carrier layer and the C2H3O+ (from the carrier layer) on the sticky layer has been effectively removed to a base level. Finally, in step 4, a high lateral resolution secondary ion image is obtained by rotating the sample so that the crosssection faces the Bi3+2 ion beam. Since the Ar-GCIB sputtering in step 3 produces a bevel cut at 45° the cross-section surface is now automatically normal to the Bi3+2 ion beam so that the spatial resolution is optimal. The SIMS image of the Scotch tape interfaces using the in situ mask method is shown in Figure 3 in comparison to the 3D SIMS data using the normal sputter depth profiling approach. Also shown are the corresponding intensity depth profiles for the peaks characteristic of the carrier layer, sticky layer and PET substrate. This serves to highlight important differences in the two methods. The in situ method preserves the flat and sharp interface between the carrier and sticky layer, which is apparently rough in the 3D SIMS image and conversely for the top surface of the carrier layer. Furthermore, the thicknesses of the carrier layer and sticky layer are 40 and 20 μm, respectively, which are accurately represented in the in situ mask method but are strongly distorted in the 3D SIMS data owing to the much higher sputtering rate of the sticky layer. This gives incorrect apparent thicknesses for the carrier layer of (45 μm) and of the sticky layer (15 μm). The thin release layer on the uppermost surface was not observed as this was removed when applying the FIB damage mask. An important advantage of the in situ mask method that is not demonstrated for this sample is the ability to observe voids correctly. These are obviously lost in the sputter depth profiling method. It is clear that the in situ mask method provides a more accurate representation of the Scotch tape interface compared with typical SIMS sputter depth profiling methods. In addition to the Scotch tape sample, we have applied the method to a multilayer polymer film used as a food package material. This successfully exposed the interface buried approximately 70 μm deep inside the film. The data is shown in the Supporting Information. This result helps demonstrate the wider applicability of our method in industrial analyses, e.g., for observations of defects, voids, impurities, and segregation of components at the interface. The in situ method allows arbitrary definition of the mask shape (including multiple masks) and the location of the mask can be placed precisely using computer control of the ion beams and the registration systems in modern SIMS instruments. The method is therefore more adaptable to automated defect analysis compared with manual mounting of masks. Furthermore, the in situ method will be superior for topographic surfaces and does not introduce any contamination. The concept is similar to providing a thin inorganic layer on the sample surface using a gas injection system, which is used for avoiding excessive damage during cross-sectioning by the FIB.17 This method is mainly applied in electron microscopic analysis and potentially useful for SIMS analysis as well.

Figure 2. Step 3, cleanup of the cross-section using low energy GCIB sputtering. Secondary ion images of C2H3O+ characteristic of the carrier layer (green) and C4H9+ characteristic of the sticky layer (red) with (a) before and (b) after 3000 s of sputtering. (c) The secondary ion intensities for C2H3O+ and C4H9+, normalized to the total intensity, from the carrier and sticky layers, respectively, following step 2 and then every 600 s cycle of cleaning.



characteristic of the carrier layer (upper layer) and is shown in green and the C4H9+ is characteristic of the sticky layer. It is clear in Figure 2a that immediately following step 3 the surface is covered with redeposited material and the interface is unclear. After 3000 s of gentle Ar-GCIB sputter cleaning this unwanted deposit is removed. In Figure 2c we show the intensities of these ions for the carrier layer and the sticky layer

CONCLUSION We have demonstrated a novel method for the cross-sectioning of organic materials without embedding or using a cryomicrotome. The method uses a gallium FIB to define a mask region of arbitrary shape that can be precisely located using the 4783

DOI: 10.1021/acs.analchem.7b00511 Anal. Chem. 2017, 89, 4781−4785

Letter

Analytical Chemistry

Figure 3. Comparison of SIMS analysis of the interfaces in Scotch tape using the in situ mask method (a and b) and traditional SIMS sputter depth profiling (c and d). Characteristic ions are shown for the carrier layer, C2H3O+ (green), sticky layer, C4H9+ (red), and PET substrate, C7H4O+ (blue). (a) Secondary ion image of cross-section, (b) intensity profile from region of interest defined by the yellow box in part a with calibrated depth scale from optical profilometry, (c) SIMS image reconstructed from sputter depth profile data, and (d) intensity depth profile from part c with depth scale assuming a constant sputtering rate and using the same optical profilometry data. The spike at the start shows existence of a thin release layer which was lost with the FIB mask method.

ORCID

computer control and the stage registration system. The mask is an area of damaged organic material that provides a protective cap to the underlying material owing to a sputtering yield several orders of magnitude lower for Ar-GCIB sputtering. A 30 keV Ga+ dose of 3.27 × 1018 ions/cm2 is used to define the mask. Other monatomic ions such as Bi+ could also be used. For the samples used in this study, the workflow takes approximately 2 h to finish. In comparison, embedding of samples in resin can take anywhere between 2 h and a day depending on the solidification time. Therefore, the proposed method is beneficial for industrial analysis of a variety of samples. It is also suited to automated analysis, which could have important benefits.



Ichiro Mihara: 0000-0001-6244-8642 Rasmus Havelund: 0000-0001-7316-9761 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work forms part of the Strategic Capability Programme (Project “3D nanoSIMS”) of the National Measurement System of the UK Department of Business, Energy and Industrial Strategy and the EMPIR Programme (Project “3DMetChemIT”). The EMPIR Programme is cofinanced by the Participating States and from the European Union’s Horizon 2020 Research and Innovation Programme.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.7b00511. Comparison of in situ mask method with no mask (using the edge of the sputter crater); optimization of the FIB dose to generate the in situ mask; and application of in situ mask method to multilayered film used as a food package material (PDF)



REFERENCES

(1) Vercammen, Y.; De Mondt, R.; Van Luppen, J.; Vangaever, F.; Van Vaeck, L. Anal. Bioanal. Chem. 2010, 396 (8), 2943−2954. (2) Plummer, H. K. Microsc. Microanal. 1997, 3 (3), 239−260. (3) Bailey, J.; Havelund, R.; Shard, A. G.; Gilmore, I. S.; Alexander, M. R.; Sharp, J. S.; Scurr, D. J. ACS Appl. Mater. Interfaces 2015, 7 (4), 2654−2659. (4) Cumpson, P. J.; Portoles, J. F.; Sano, N. J. Vac. Sci. Technol., A 2013, 31 (2), 020605. (5) Robinson, M. A.; Graham, D. J.; Castner, D. G. Anal. Chem. 2012, 84 (11), 4880−4885. (6) Barnes, J.-P.; Djomeni, L.; Minoret, S.; Mourier, T.; Fabbri, J.-M.; Audoit, G.; Fadloun, S. J. Vac. Sci. Technol., B: Nanotechnol. Microelectron.: Mater., Process., Meas., Phenom. 2016, 34 (3), 03H137. (7) Iida, S.; Miyayama, T.; Fisher, G. L.; Hammond, J. S.; Bryan, S. R.; Sanada, N. Surf. Interface Anal. 2014, 46, 83−86.

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. 4784

DOI: 10.1021/acs.analchem.7b00511 Anal. Chem. 2017, 89, 4781−4785

Letter

Analytical Chemistry (8) Seah, M. P. J. Phys. Chem. C 2013, 117 (24), 12622−12632. (9) Wucher, A.; Cheng, J.; Winograd, N. Anal. Chem. 2007, 79 (15), 5529−5539. (10) Wucher, A.; Cheng, J.; Zheng, L.; Winograd, N. Anal. Bioanal. Chem. 2009, 393 (8), 1835−1842. (11) Vanbellingen, Q. P.; Elie, N.; Eller, M. J.; Della-Negra, S.; Touboul, D.; Brunelle, A. Rapid Commun. Mass Spectrom. 2015, 29 (13), 1187−1195. (12) Claus, T. K.; Richter, B.; Hahn, V.; Welle, A.; Kayser, S.; Wegener, M.; Bastmeyer, M.; Delaittre, G.; Barner-Kowollik, C. Angew. Chem., Int. Ed. 2016, 55 (11), 3817−3822. (13) Lee, J. L. S.; Gilmore, I. S.; Seah, M. P.; Levick, A. P.; Shard, A. G. Surf. Interface Anal. 2012, 44 (2), 238−245. (14) Mahoney, C. M. Mass Spectrom. Rev. 2010, 29 (2), 247−293. (15) Gilmore, I. S.; Seah, M. P. Surf. Interface Anal. 1996, 24 (11), 746−762. (16) Lorenz, M.; Shard, A. G.; Counsell, J. D. P.; Hutton, S.; Gilmore, I. S. J. Phys. Chem. C 2016, 120 (44), 25317−25327. (17) Hayles, M. F.; Stokes, D. J.; Phifer, D.; Findlay, K. C. J. Microsc. 2007, 226 (3), 263−269.

4785

DOI: 10.1021/acs.analchem.7b00511 Anal. Chem. 2017, 89, 4781−4785