Fabrication of sharp and smooth Ag tips is crucial in optical scanning probe microscope experiments. To ensure reproducible tip profiles, the polishing process is fully automated using a closed-loop laminar flow system to deliver the electrolytic solution to moving electrodes mounted on a motorized translational stage. The repetitive translational motion is controlled precisely on the μm scale with a stepper motor and screw-thread mechanism. The automated setup allows reproducible control over the tip profile and improves smoothness and sharpness of tips (radius 27 ± 18 nm), as measured by ultrafast field emission.

Plasmonically active, sharp noble metal tips are essential in experiments such as photon emission at a scanning tunneling microscope (STM) junction,1 ultrafast electron emission,2,3 tip enhanced Raman spectroscopy (TERS),4,5 and superfocusing of propagating plasmons.6–8 Among noble metals, Ag and Au are the most pertinent because of their low dielectric losses. Au tips are more commonly used because of their inertness in air, but Ag tips are favored over Au tips in vacuum because they are less lossy in the relevant spectral range; and as such, plasmon propagation lengths on Ag are significantly longer.9 In practice, silver tips suitable for the aforementioned applications are more challenging to prepare than regular STM tips. Here, we present a fully automated method for reliably producing such Ag tips, which we verify through the stringent quality test of the stability of their field emission under ultrafast laser irradiation.

Although Ag tips are attainable by a single-step etching,10,11 it is challenging to control the smoothness of the etched surface in this process; Ag is known to easily roughen. Scattering losses due to surface roughness defeat the advantageous optical properties, and asperities develop local hotspots other than the tip apex compromising applications such as TERS. A second stage of electrochemical polishing12 is necessary when smooth surfaces matter in addition to sharpness. Polishing is carried out by repetitively moving an electrolytic solution membrane sustained by a small loop, in a back and forth motion.13–16 The procedure is labor intensive, requiring replenishing of the membrane because Ag oxides quickly build up inside the loop. Redeposition of the insoluble silver oxides on the surface is a principal mechanism of surface roughening. We overcome this problem in the constant flow automated electropolishing apparatus, which we describe here.

We employ a two-stage electrochemical process. In the first stage, a rough etch is performed in a Petri dish filled with a solution of volume ratio 1:1:1 NH4OH (28%):H2O2 (30%):CH3OH, according to a well-established procedure.17 We prefer this relatively benign basic solution over toxic acids.18,19 A 1 in. diameter platinum wire loop cathode is placed concentrically with the tip anode at the center, and ∼15 V dc bias is applied between electrodes. A samarium-cobalt magnet (McMaster-Carr, part # 5768K24) placed underneath the Petri dish improves the azimuthal uniformity of the tip surface.20 

The second stage is the repetitive process of electrochemical polishing to smoothen the surface and control the taper. We use a solution of volume ratio 6:1 NH4OH:CH3OH to form a stable laminar flow. We avoid using H2O2 because it leads to bubbling and disturbs laminarity. A key novelty in the present setup is the use of the wire-guided constant electrolytic flow system illustrated in Fig. 1, following a design implemented by Tauber et al. in a very different context.21 As the ring passes back and forth over the tip, a dc bias ∼10 V is applied. This dynamic zone-etching process13 is limited to the area under the solution membrane.

FIG. 1.

Electrochemical polishing setup. (a) Overall schematic of the automated etching system. (b) Lower right: (1) 1/4 in. barbed to 1/8 NPT adapter; (2) 1/8 NPT to 1/8 in. Swagelok; (3) 1/8 OD stainless steel tube. Upper left: magnified view of the lower right; (4) 0.25 mm Pt wire loop; (5) lens paper ribbon as a drain guide.

FIG. 1.

Electrochemical polishing setup. (a) Overall schematic of the automated etching system. (b) Lower right: (1) 1/4 in. barbed to 1/8 NPT adapter; (2) 1/8 NPT to 1/8 in. Swagelok; (3) 1/8 OD stainless steel tube. Upper left: magnified view of the lower right; (4) 0.25 mm Pt wire loop; (5) lens paper ribbon as a drain guide.

Close modal

Repeated passes eventually form a neck, which becomes the breaking point of the wire. Etching is terminated immediately upon detachment of the excess wire.

The overall schematic of the hardware is shown in Fig. 1. The tip is horizontally mounted in a socket (Digi-Key A24862-ND) attached to a standard optical post and based on an XYZ translational stage. A video microscope (Meiji VM-1V) displays the image of the tip apex on the computer monitor. The combination of a long working distance objective lens (20 mm, MA870 Plan Epi Metallurgical 5×) and the projection of a CCD camera (Moticam 1000) image to the monitor provides 320× magnification.

The reservoir of etching solution is connected via flexible Tygon tubing, through a valve and a Swagelok™ adapter to a 1 in. long stainless steel tube. The tube is crimped using a vise at one end to hold the platinum (Alfa Aesar, 0.25 mm) wire cathode. A 1.5 mm wide, 0.23 mm thick stainless steel sheet was inserted with the Pt loop and they were crimped together to open a slit-like nozzle. The resultant slit forms a stable laminar flow. This Pt cathode is roughly ring shaped, with a diameter of 0.2 in., as illustrated in Fig. 1(b). This arrangement allows gravitational flow to form a thin membrane suspended on the wire, and a lens paper ribbon (0.25 in. wide, 1.5 in. long) hanging from it guides the runoff to the sump and prevents droplets forming near the membrane. The cathode ring is angled at 45° to give the objective lens access to image the etching process in real time. The flow rate from the reservoir is controlled by a Polytetrafluoroethylene (PTFE) needle valve (Omega, FVLT101). The sump, a 3 in. diameter Petri dish, holds the runoff, and a peristaltic pump (Cole-Parmer, 77200-50) circulates it back into the reservoir. The cathode assembly is mounted on a custom motion stage. The leads of the dc voltage supply are clipped on (to the tube holding the loop and the post holding the tip). Once the membrane is self-established by fully opening the needle valve, its thickness and flow rate can be adjusted by the combination of the needle valve and the height of the reservoir.21 

Figure 2 shows the electronic and mechanical schematic of the system. The computer uses a motion control card (National Instruments PCI-7330) to communicate with the multi-axis motion interface (National Instruments UMI-7664).

FIG. 2.

Schematic of the automated etching system. (a) Arrangement of electronic and mechanical components. (b) Exploded view drawing of the custom motion stage. The upper left detail illustrates the connection between the motor and the threaded rod, while the bottom right shows the assembled unit.

FIG. 2.

Schematic of the automated etching system. (a) Arrangement of electronic and mechanical components. (b) Exploded view drawing of the custom motion stage. The upper left detail illustrates the connection between the motor and the threaded rod, while the bottom right shows the assembled unit.

Close modal

The interface is then connected to the driver (Danaher P70530) of the stepper motor (Danaher CTP10ELF10MAA00).

The computer communicates via a universal serial bus (USB) port with a relay switch (National Instruments USB-6525) placed between the positive terminal of the power supply and the tip, to turn the applied voltage on/off. The negative terminal is directly connected to the platinum ring. The motion stage we designed and built is detailed in Fig. 2(b). The rugged design is to provide a maintenance-free system in a chemically toxic environment. The key is a screw-thread mechanism that converts the rotational motion of the stepper motor to translational motion without introducing gears, springs, or bearings. The parts shown in the exploded view are, from top right to bottom left: stepper motor, aluminum coupling plate to which the motor is attached, a rubber gasket, a threaded rod, a brass shaft with two brass wings, and the main aluminum body and support. The threaded rod is a left-handed 1/2-10 thread, such that “forward” rotation of the motor results in forward motion of the shaft. The combination of the thread pitch and 1.8° minimum step angle leads to a minimum translation distance of 12.5 μm. The back face of the threaded rod has a hole bored to fit the shaft of the motor. The main shaft is a 1-1/2 in. diameter brass cylinder with a 1/2-10 left-handed tapped hole through the center axis. The midpoint of the shaft has a flattened strip on opposite sides, where two small brass blocks are bolted to protrude beyond the diameter of the shaft. These wings fit into the side slots in the main body and prevent the shaft from rotating, limiting its motion to translation. The front face of the shaft has a 1/4-20 tapped hole for mounting a standard optical post, onto which the platinum ring is mounted. The main body is a solid block of aluminum through which is bored a 1-1/2 in. diameter hole with side slots for the wings of the shaft.

As the motor turns forward, the threaded rod unscrews from the main shaft. Because the rod is fixed, the shaft is pushed forward. When the motor is reversed, the translation of the shaft is reversed. In this way the platinum ring is advanced or retracted relative to the tip being etched.

The etching procedure is automated through software control. In the initial stage, etching is cycled over a travel range of ∼2 mm until a neck forms (see optical microscope images in Fig. 3).

FIG. 3.

Evolution of tip profiles. [(a)–(d)] A series of images showing the evolution of a neck on the tip. [(e) and (f)] Scanning electron microscope (SEM) images of a tip prepared by (e) the first etching stage only with an inset showing the magnified apex, and (f) a tip that has been polished with an inset showing the tip apex with a radius of curvature of ∼50 nm.

FIG. 3.

Evolution of tip profiles. [(a)–(d)] A series of images showing the evolution of a neck on the tip. [(e) and (f)] Scanning electron microscope (SEM) images of a tip prepared by (e) the first etching stage only with an inset showing the magnified apex, and (f) a tip that has been polished with an inset showing the tip apex with a radius of curvature of ∼50 nm.

Close modal

When the neck thins, we start the final phase where the voltage is applied only when the loop is moving away from the tip. This unidirectional etching polishes the surface by etching from the shank to the apex. The taper of the apex is determined by the polishing range. The sharpness is improved by minimizing the time spent at the apex, which is limited by the thickness of the solution film and the speed of motion of the ring. When the neck approaches the breaking point (<10% diameter of surrounding apex, Fig. 3(c)), the cycling mode is stopped, and motion continues with single strokes to ensure that etching does not continue after break-off. The final result is shown in Fig. 3(d).

Figures 3(e) and 3(f) show a comparison between a tip prepared by only first etching and a tip polished by the automated setup. The 1st-stage etched tip in Fig. 3(e) exhibits facets due to preferential etching along particular crystal axes. Upon polishing, the sharpness and smoothness are greatly improved as demonstrated in Fig. 3(f). Out of 10 polished tips, we obtained an average radius of curvature of 27 ± 18 nm.

To characterize the polishing effect, we perform femtosecond laser-induced field emission (FE) tests in a high vacuum chamber with a base pressure of 7 × 10−6 Torr. A 20 fs pulse train is focused on the tip apex as shown in Fig. 4(a), and the time trace of field emission current is recorded. The fs laser FE test of tips is more stringent than the STM image scan or tunneling induced photon emission,1 because FE requires a sharp apex under the laser focal spot (∼5 μm diameter), whereas the STM concerns only a few atoms at the apex for scanning, and nm scale volume for photon emission. Representative FE current time traces of a 1st-stage etched and polished tip over 700 s upon 15 mW laser irradiation are compared in Fig. 4(b). With the same given laser power, the polished tip produces a much higher and steady current (average 27 pA) compared to the 1st-stage etched tip which produces a small initial FE current (∼2 pA) followed by a decay to half in amplitude within 100 s. Four other 1st-stage etched tips under the same examination exhibited much shorter decay time, whereas all four tested polished tips produced stable current. We explain the difference as follows. The larger current from the polished tip is attributed to the sharper apex leading to stronger field enhancement, while the stability of the current is attributed to the smooth finish of the surface. By contrast, the smaller current from the 1st-stage etched tip indicates a blunter apex. The instability of the field emission current is ascribed to asperities, which evolve in shape under the strongly localized plasmon fields.22 The current decays as sharp asperities melt away.

FIG. 4.

Femtosecond laser-induced electron emission test. (a) Experimental setup. A 15 mW, 20 fs laser pulse train polarized along the tip axis is focused at the tip apex. (b) Comparison of time traces between a 1st-stage etched and polished tip.

FIG. 4.

Femtosecond laser-induced electron emission test. (a) Experimental setup. A 15 mW, 20 fs laser pulse train polarized along the tip axis is focused at the tip apex. (b) Comparison of time traces between a 1st-stage etched and polished tip.

Close modal

Automation of both the electrode motion and the flow of etching solution has dramatically increased both the speed and reproducibility of sharp tip preparation of an otherwise labor intensive procedure. Polished silver tips prepared by the automated setup are demonstrated to perform better than 1st-stage etched tips in fs laser-induced electron emission.

While we have carried out the demonstrations on the more challenging preparation of silver tips, the same apparatus with different solutions can be applied to etch and polish other materials, such as tungsten.

Funding was provided by the National Science Foundation Center for Chemical Innovation dedicated to Chemistry at the Space-Time Limit (Grant No. CHE-082913). S.S.S. is grateful to the summer undergraduate research program (SURP) for the 2009 grant, and A.R. is grateful for his NSF graduate research fellowship (Grant No. DGE-0808392).

1.
R.
Berndt
,
J.
Gimzewski
, and
P.
Johansson
,
Phys. Rev. Lett.
67
,
3796
(
1991
).
2.
P.
Hommelhoff
,
Y.
Sortais
,
A.
Aghajani-Talesh
, and
M. A.
Kasevich
,
Phys. Rev. Lett.
96
,
077401
(
2006
).
3.
J.
Lee
,
S. M.
Perdue
,
D.
Whitmore
, and
V. A.
Apkarian
,
J. Chem. Phys.
133
,
104706
(
2010
).
4.
W.
Zhang
,
B. S.
Yeo
,
T.
Schmid
, and
R.
Zenobi
,
J. Phys. Chem. C
111
,
1733
(
2007
).
5.
C. C.
Neacsu
,
S.
Berweger
, and
M. B.
Raschke
,
NanoBiotechnology
3
,
172
(
2007
).
6.
C.
Ropers
,
C. C.
Neacsu
,
T.
Elsaesser
,
M.
Albrecht
,
M. B.
Raschke
, and
C.
Lienau
,
Nano Lett.
7
,
2784
(
2007
).
7.
A. J.
Babadjanyan
,
N. L.
Margaryan
, and
Kh. V.
Nerkararyan
,
J. Appl. Phys.
87
,
3785
(
2000
).
8.
M. W.
Vogel
and
D. K.
Gramotnev
,
J. Appl. Phys.
107
,
044303
(
2010
).
9.
H.
Raether
,
Surface Plasmons on Smooth and Rough Surfaces and on Gratings
(
Springer
,
New York
,
1988
), p.
7
.
10.
K.
Dickmann
,
F.
Demming
, and
J.
Jersch
,
Rev. Sci. Instrum.
67
,
845
(
1996
).
11.
C.
Zhang
,
B.
Gao
,
L. G.
Chen
,
Q. S.
Meng
,
H.
Yang
,
R.
Zhang
,
X.
Tao
,
H. Y.
Gao
,
Y.
Liao
, and
Z. C.
Dong
,
Rev. Sci. Instrum.
82
,
083101
(
2011
).
12.
Traditionally polishing and etching were differentiated by the current density (Ref. 13). Here, we use the terms in the mechanical sense; static and dynamic electrochemical processes are called etching and polishing, respectively.
13.
A. J.
Melmed
,
J. Vac. Sci. Technol. A
2
,
1388
(
1984
).
14.
A. A.
Gorbunov
,
B.
Wolf
, and
J.
Edelmann
,
Rev. Sci. Instrum.
64
,
2393
(
1993
).
15.
B. W.
Krakauer
,
J. G.
Hu
,
S.-M.
Kuo
,
R. L.
Mallick
,
A.
Seki
,
D. N.
Seidman
,
J. P.
Baker
, and
R.
Loyd
,
Rev. Sci. Instrum.
61
,
3390
(
1990
).
16.
B. W.
Krakauer
and
D. N.
Seidman
,
Rev. Sci. Instrum.
63
,
4071
(
1992
).
17.
F.
Okamoto
, U.S. patent 3860423 (14 January
1975
).
18.
B.
Ren
,
G.
Picardi
, and
B.
Pettinger
,
Rev. Sci. Instrum.
75
,
837
(
2004
).
19.
J. S.
Lloyd
,
A.
Williams
,
R. H.
Rickman
,
A.
McCowen
, and
P. R.
Dunstan
,
Appl. Phys. Lett.
99
,
143108
(
2011
).
20.
Ungdon
Ham
, private communication (
2010
).
21.
M. J.
Tauber
,
R. A.
Mathies
,
X.
Chen
, and
S. E.
Bradforth
,
Rev. Sci. Instrum.
74
,
4958
(
2003
).
22.
J.
Li
,
R.
Berndt
, and
W.
Schneider
,
Phys. Rev. Lett.
76
,
1888
(
1996
).