Feasibility of a New Atom Probe Specimen Preparation Method Using a Focused Ion Beam

Specimens for three-dimensional atom probe (3DAP) analysis must be needle-shaped and the apex of specimen is primarily fabricated by focused ion beam (FIB). In the specimen preparation by FIB, the gallium ions are generally implanted into the surface region of the specimen during the irradiations of the ions. Therefore, the surface structure of the specimen is disarranged and amorphous. This phenomenon makes it difficult to reconstruct a three-dimensional image of the specimen. The implantation is caused by irradiating the apex of the specimen with FIB. In this study, we propose a new specimen preparation method in which FIB is irradiated from behind the needle specimen due to avoids the implantation of gallium ions. The assembly in this method was installed in the FIB instrument. The specimens were fabricated by means of the conventional and proposed methods and analyzed by atom probe in our laboratory. It was shown that not only the gallium implantation but also the rupture of specimen were inhibited by the proposed method. [DOI: 10.1380/ejssnt.2009.863]


I. INTRODUCTION
The scale of electronic devices has been getting smaller and smaller.For example, the commercialized devices are designed within 50-nm rules now, and are expected to be designed within 1-nm rules in the next decade.In the analyses of such small devices, the performance of conventional analytical methods, including secondary ion mass spectrometry and transmission electron microscope, is limited in spatial resolution and elemental analysis.The three-dimensional atom probe (3DAP) is one of the most powerful methods for three-dimensional and elemental analyses on an atomic scale [1].However, there are some problems to apply this method to electric devices [2].
In 3DAP analysis, specimens must be needle-shaped and the apex of specimen is primarily fabricated using fo- * Corresponding author: mayama@iis.u-tokyo.ac.jp cused ion beam (FIB) methods [3,4].Two methods are mainly used for the specimen preparation.One method is annular milling, in which a specimen is milled circumferentially [5].The other method is a linear cut across the apex region of the specimen [5].Ion milling with high-energy gallium ions generally causes implantation of the ions into the surface region of the specimen [6].The gallium ions implant into the apex of specimen in the both methods.The surface structure of the specimen is disarranged and amorphous [7].To decrease the implantation, some specimen preparation methods were reported [6].However, slight implantation still occurs.In these methods, the FIB was irradiated from the direction of the specimen apex.The implantation might be caused by irradiation of circumference of the FIB to the apex of the specimen.
In this study, we propose a new specimen preparation method in which FIB is irradiated from behind the needle specimen due to avoids the implantation.The circumference of FIB irradiates to the side of the specimen, and the apex of the specimen is not irradiated.The assembly in this method was installed in the FIB instrument.The  specimens were fabricated by means of the conventional (annular milling) and proposed methods and analyzed by atom probe in our laboratory.

A. Instrumentation
The assembly in this method in which FIB is irradiated from behind the needle specimen is shown in Fig. 1.A commercial FIB instrument (SMI-3050SE, SII NanoTechnology Inc.) was chosen as the base instrument.To fabricate the specimen like a circular cone, a system to rotate the specimen is necessary.In this study, a small stepping motor was used from three circumstances, 1) the diameter of motor must be <10 mm due to the spatial limitations of the FIB instrument, 2) the rotational wobble of motor shaft must be <100 µm due to the scope area of the FIB, and 3) the rotation angle must be precisely controllable.The specimen stage holder (Fig. 1) was modified to enable rotation in increments of about 1 degree.

B. Procedure
The specimen and the stepping motor shaft were connected in a straight line by using a chucking jig, then fixed on the specimen stage holder as shown in Fig. 1.The rotational wobble of the specimen apex was reduced to <100 µm by adjusting the chucking.The stage holder was put into the FIB instrument, the specimen was set perpendicularly against the beam axis of the FIB, and the stage holder was tilted to 45-60 degrees.FIB etches the apex from behind the specimen.The stepping motor was then rotated by the proper angle and the specimen was etched by the FIB.This cycle was continued until the motor had rotated 360 degrees.

C. Specimens
Two specimens were prepared from a 99.95 % pure tungsten wire (diameter 0.1 mm; NILACO #W-461167).Tungsten wire was cut to 10-15 mm in length and the rear edge of the wire was fixed in the Cu-tube.The specimens were electropolished to an end radius of typically less than <100 nm using a solution of 5 % sodium hydroxide in pure water.Then, the specimen apexes were fabricated by the FIB.
The apex of specimen A was fabricated by the conventional annular milling method [3].That of specimen B was fabricated by the proposed method.The applied voltage of FIB in both cases was 30 kV.In the preparation of specimen B, the stage holder was tilted to 45 degrees and the stepping motor was rotated by 36 angles.

D. Atom probe analyses
Atom probe analyses for both specimens were performed in a homemade atom probe instrument with a vacuum of 3×10 −8 Pa triggered with a voltage pulse [8].The pulse voltage was 250 V, pulse repetition rate was 2 kHz, and pulse width was <30 ns.The aperture size of the counter electrode was ∼200 µm.The analyses were performed at room temperature.

III. RESULTS AND DISCUSSIONS
Figure 2 shows a scanning electron microscope (SEM) image of the apex of specimen B. The image shows that the radius of curvature at the apex was <50 nm and specimen B was suitable for 3DAP analysis.The taper angle was ∼90 degrees (Fig. 2), which was ideal angle calculated from the result of our simulation of endurance against stress by the electric field at the apex [9].In contrast, the taper angle of specimen A fabricated by the annular milling method was far smaller [3].Field evaporation was observed at the applied voltage of 2.6 kV in the analysis of specimen A and at 2.7 kV for specimen B. The voltage of field evaporation depends on the radius of curvature at the apex [1], so the two specimens had almost equal radii.
Figure 3 shows the mass spectrum of specimen A, fabricated by the annular milling method.W 3+ , W 4+ , and Ga 3+ were detected (Ga 3+ : 16, W 4+ : 24, W 3+ : 223 at counts). Figure 4 shows the mass spectrum of specimen B, fabricated by the proposed method.W 3+ and W 4+ were detected (W 4+ : 112, W 3+ : 735).The ratio of W 4+ /W 3+ in the analysis of specimen A was ∼11 % and that of specimen B was ∼15 %.The ratio of detected ions depends on http://www.sssj.org/ejssnt(J-Stage: http://www.jstage.jst.go.jp/browse/ejssnt/) e-Journal of Surface Science and Nanotechnology the electric field at the specimen apex [10], so almost the same electric fields arose at the apexes of two specimens.
In the mass spectrum of specimen A (Fig. 3), the ratio of gallium ions to all detected ions was ∼6 %.Although W 4+ and Ga 3+ in the analysis of specimen A are detected at about the same counts, gallium ions in specimen B (Fig. 4) were not detected.Even though all signals in the mass to charge ratio between 23 and 24 were Ga 3+ , the ratio of gallium ions to all detected ions was only ∼1 % (Ga 3+ : 7, W 4+ : 112, W 3+ : 735).In the proposed method, the implantation of gallium ions decreased.
Total counts of detected ions differ in the analyses of two specimens.In specimen A, the applied voltage to the specimen was gradually increased to maintain a constant counting rate of detected ions.At 2.8 kV, no ions were detected.The specimen A had a small taper angle and had low endurance against stress by the electric field [9].Then, the specimen ruptured.Specimen B had a large taper angle and its radius of curvature increased as field evaporation proceeded.To maintain a constant counting rate, the applied voltage was increased from 2.7 to 3.5 kV much more rapidly than that in specimen A. However, specimen B did not rupture.From these results, it was demonstrated that the specimen with wider taper angle had greater endurance against stress by the electric field.Therefore, it was shown that not only the gallium implantation but also the rupture of specimen were inhibited by the proposed method.

IV. CONCLUSION
To avoid the implantation of gallium ions, we propose a new specimen preparation method in the 3DAP in which FIB is irradiated from behind the needle specimen.The assembly in this method was installed in the FIB instrument.To evaluate the implantation of gallium ions in the specimens, two specimens were prepared by means of the conventional and proposed methods and analyzed by atom probe in our laboratory.As a result, it was shown that not only the gallium implantation but also the rupture of specimen were inhibited by the proposed method.

FIG. 1 :
FIG. 1: Assembly to fabricate a specimen by the proposed method.

FIG. 2 :
FIG. 2: SEM image of specimen B fabricated by the proposed method.

FIG. 3 :
FIG. 3: Mass spectrum of specimen A fabricated by the annular milling method.