September_AMP

A D V A N C E D M A T E R I A L S & P R O C E S S E S | S E P T E M B E R 2 0 1 5

3 0

performance. Copper phthalocyanine

(CuPc) is of significant interest as an or-

ganometallic semiconductor and is the

basis for the device under study.

Using a depth profiling technique,

XPS can monitor the composition of

a coated or layered material to great-

er depths (hundreds of nm or even µm

scales). Traditional ion sputtering depth

profiling techniques cause significant

damage to the subsurface of “soft” or-

ganic materials, which is visible in XPS

data as a loss of structure and chemical

information. The new MAGCIS source

addresses this problem. The gentle,

shallow sputtering from argon cluster

ions does not cause subsurface damage,

so XPS data preserve chemical state in-

formation throughout the entire profile.

MAGCIS can also generate traditional

monatomic ion beams for profiling in-

organic materials, so it is possible to

profile through an organic layer with the

cluster ions, then switch to monatomic

ions to profile into a harder substrate.

The K-Alpha

XPS instrument, fit-

ted with the MAGCIS cluster ion source,

was used to acquire depth profiles from

the CuPc FET, and was profiled with

4 keV argon cluster ions, with an aver-

age cluster size of 2000. The source was

then switched to monatomic ion gener-

ation to profile through the SiO

isola-

tion layer into the silicon substrate. The

two profiles were combined to produce

a single depth profile through the entire

organometallic/inorganic stack.

In this analysis, the spectra are

similar, but the device sample has more

hydrocarbon signal from surface con-

taminants at 285 eV. This confirms that

the surface chemistry of the organic

layer approximates CuPc. MAGCIS clus-

ter ion sputtering successfully removed

material without causing subsurface

damage, so the XPS spectrum resem-

bles a pure material. This is a primary

strength of the cluster ion technique,

which allows depth profile analysis of

organic materials without visible dam-

age in XPS data. The spectral quantifi-

cation in Table 1 confirms that the ex-

posed CuPc is an excellent match to the

expected composition.

By sputtering through the CuPc

with cluster ions, and through the SiO

with monatomic ions, building up a

profile of the entire device stack is pos-

sible (Fig. 2).

CLEANING METAL OXIDES

The benefits of cluster ion sources

are not limited to depth profiling exper-

iments; they can also be useful for more

routine analysis. For example, accu-

rately quantifying the chemical compo-

sitions of metal oxide samples with XPS

can be hindered by adventitious car-

bon contamination, typically present

on the surface. Gas cluster ions do not

penetrate the oxide surface, preserving

chemical state information while still

cleaning contamination from a sample.

With the ability to provide both cluster

ions and monatomic ions, the MAGCIS

source offers versatility and conve-

nience for a range of sample types.

The K-Alpha

XPS instrument was

used to acquire spectra froma sample of

tantalumpentoxide. Spectra were taken

at three different areas following differ-

ent cleaning processes. In the first area,

spectra were taken in an as-received

state with no cleaning performed; in

the second position, the sample was

sputtered with 200-eV monatomic ions

before analysis; and at the third posi-

tion, the sample was analyzed following

sputtering with single-ion argon clus-

ters. For this experiment, cluster beam

energy was 4 keV with a cluster size of

1000 atoms. Survey spectra (wide scans

covering the entire elemental range)

were also collected from uncleansed

areas, as well as the area that had been

sputtered with the cluster ion beam.

Fig. 1 —

Schematic of copper phthalocyanine (CuPc)

organic FET.

TABLE 1 —QUANTIFICATION OF

MAGCIS-CLEANED CuPc

Expected

atomic %

Observed

atomic %

78.0

78.6

19.5

19.4

2.4

1.9

Fig. 2 —

MAGCIS profile of organic FET.