Using a Surface Analysis System to Determine Layer Structures of an Organic FET Device

A copper phthalocyanine-based organic material, designed for use in FETs, was analyzed using a Thermo Scientific™ Nexsa™ Surface Analysis System, equipped with the Thermo Scientific™ MAGCIS™ dual mode ion source. The organic layer was profiled using argon clusters, which ensured the retention of chemical information in the XPS data, followed by profiling of the silicon substrate using monoatomic argon.

Introduction

The pressure to produce increasingly thin, flexible and economical electronics has resulted in the design of novel organic microelectronics, such as field effect transistors (FETs). Initial designs of FETs mostly used aromatic organic semiconductors, whereas modern methods use organometallics for performance reasons.

Copper phthalocyanine (CuPc, Figure 1) is one such organometallic semiconductor, and is present in the device studied in the following research.

Figure 1: Copper phthalocyanine

The thickness and the composition of the different organometallic and organic layers in organic FETs play an important role in the way in which a device performs. X-ray photoelectron spectroscopy (XPS) can be used to obtain information on a sample’s chemical composition from its surface, to a depth of 5 – 10 nm, making it perfect for surface analysis. If combined with a depth profiling technique XPS can be used to analyze coated surfaces, or layered systems, to even greater depths that can reach up to a μm scale.

Traditional depth profiling methods, i.e. conventional ion sputtering, can result in the subsurface of organic materials, which tend to be softer, being damaged. This damage translates into XPS data that contains less chemical and structural information. Thermo Scientific’s MAGCIS argon cluster ion source solves this problem.

The argon cluster ions provide a method of sputtering which is shallower and gentler, meaning subsurface features are not damaged; allowing the XPS data to retain chemical state information throughout the profile. The source used by MAGCIS is also capable of producing the monatomic ion beams found in conventional ion sources, which can be used to profile harder substrates such as silicon. This means MAGCIS can be used to analyze both soft and hard substrates.

Method/Experiment

Depth profiles of the CuPc FET were taken using the Thermo Scientific Nexsa Surface Analysis System, equipped with the MAGCIS cluster ion source. Adequate top-surface electrical content was guaranteed by the use of a standard Nexsa sample holder to hold the CuPc sample in place.

Figure 2: Schematic of CuPc organic FET

Profiling of the CuPc sample was achieved using 4 keV argon cluster ions, which had an average cluster size of 2,000 ions each. Following this, monochromatic ions were used to profile through the oxidized surface and deeper into the silicon substrate. The two different profiles (cluster + monochromatic) were then combined to produce a single depth profile with data on the whole organometallic/inorganic layered system.

Results

Figure 3 shows a comparison between the collected XPS 1s spectra taken from the CuPc device’s surface and a reference sample. Whilst the spectra show similarities the experimental sample shows signals at 285 eV, which are indicative of surface contamination from hydrocarbons. This confirms that the surface chemistry of the organic layer approximates CuPc.

Figure 3: Sample and reference C1s spectra

A C1s spectrum of the sample following removal of 6 nm of its surface, via cluster ion sputtering, is shown in Figure 4. This spectrum is far more similar to the reference spectrum, confirming that the device’s semiconducting layer consists of pure CuPc. The purity of the CuPc analyzed also demonstrates that the MAGCIS’s cluster ion sputtering method has removed material without causing significant sub-surface damage.

Figure 4: Quantification of MAGCIS-cleaned CuPc

The fact that organic materials can be sputtered for depth profiling without damage that appears in the XPS data is a key advantage of using a cluster ion method. The spectral data shown in Table 1 confirms that the sub-surface CuPc matches well with its expected composition.

Table 1: Quantification of MAGCIS-cleaned CuPc

Deconvolution of the spectrum using peak fitting was used to demonstrate how the XPS data relates to the chemistry of the sample. The blue peaks correspond to carbon atoms bonded to nitrogen in the 5-membered rings, the red peaks to the carbon atoms in the 6-membered rings and the green peaks are the result of a complex group of loss features relating to the aromatic ring systems.

The fact that these loss features have been preserved during the analysis indicates that the cluster ion beam did not significantly damage the sample.

The use of cluster ions to sputter through the CuPc layer, followed by monoatomic ions to sputter through the SiO₂ layer, allows researchers to create a depth profile of the whole device (Figure 5). Following the removal of modifications and surface contamination the CuPc has retained its correct stoichiometry throughout the sample, as made possible using the softer method of cluster ion sputtering. The monoatomic ion sputtering of the SiO₂ also showed it had the correct stoichiometry until the bulk Si substrate is reached.

Figure 5: MAGCIS profile of organic FET

This complete quantitative profile, obtained using a single ion source, is not feasible by other means.

Summary

Themo Scientific’s Nexsa Surface Analysis System was equipped with a MAGCIS ion source to take a depth profile of an organic semiconductor layer (CuPc) on an SiO₂/Si substrate.

The MAGCIS source allowed the CuPc layer to be profiled using a cluster ion source and for the insulating SiO₂ layer to be profiled using a monoatomic ion source.

As more and more electronic devices use soft semiconductors (such as organic polymers) approaches such as this become increasingly essential for researchers who need to determine the layer structure of devices, carry out failure analysis and to gain insight into changes at interfaces.

Acknowledgments

Thermo Fisher Scientific would like to thank Prof. Dr. Dietrich R. T. Zahn, Dr. Daniel Lehmann and Iulia Korodi from NanoMA, Chemnitz University of Technology, for supplying the CuPc organic FET sample.

Produced from materials originally authored by Paul Mack from Thermo Fisher Scientific East Grinstead, West Sussex, UK.

This information has been sourced, reviewed and adapted from materials provided by Thermo Fisher Scientific – X-Ray Photoelectron Spectroscopy (XPS).

For more information on this source, please visit Thermo Fisher Scientific – X-Ray Photoelectron Spectroscopy (XPS).