Real-space visualization of atomic- and molecular-scale structures has always attracted researchers, and they welcomed the invention of scanning tunneling microscopy (STM) in 1981. STM images of many crystalline compounds revealed their atomic and molecular arrangements as well as various single-atom defects. STM applications, however, were limited due to requirement of sample conductivity.

Therefore, atomic force microscopy (AFM), which can be applied to any kinds of samples, became the leading scanning probe method. For the past 20 years AFM has been developed into a powerful characterization technique but the question of atomic-resolution in AFM is still opened. The initial AFM measurements were performed in the contact mode and the atomic-scale lattices were revealed first in the images of graphite, MoS2, BN and mica [1]. Similar observations were later reported for a large number of organic and inorganic crystals and layers [2]. These images were usually obtained using Si3N4 probes (an averaged tip size ~ 20 nm) and soft Si probes (tip size in the 10-20 nm range).

The main drawback of these images was an absence of atomic-scale defects that implies only lattice but not true atomic resolution [3]. The latter suggests a presence of single-molecular defects that, indeed, were observed in AFM images obtained in oscillatory frequency modulation (FM) and amplitude modulation (AM) modes. The images of surface lattices of Si (7x7) and InP crystals were obtained in FM studies in UHV [4-5]. Latter, true atomic-scale FM visualization of mica and polydiacetylene (PDA) crystals was achieved in air and under water [6-8].

These results were obtained by researchers which used custom instruments with optical deflection and interferometric detection systems. It also became clear that FM is not unique oscillatory mode for observation of surface lattices with defects and similar results were achieved in AM studies in air [9-10]. The related images of polydiacetylene crystal obtained with Dimension 5000 (Veeco Instruments) and Agilent 5500 (Agilent Technologies) microscopes, Figures 1a-b. In these studies different MikroMasch probes (Hi-Res-C and DP-14) were applied. The molecular-scale observations in AM, whose feedback mechanism is slower than in FM, require a low thermal drift environment. Also it is easy to achieve the molecular-scale resolution using most sharp probes if they are not damage in the engagement procedure and preserved from breaking in the low-force imaging.

For users of MultiMode and Dimension 3100 and 5000 SPM microscopes (Veeco Instruments) there are more specific recommendations for preventing sharp probes from damage:

1. Excessive tip-sample forces can be generally avoided by using softer cantilevers though there are limitations caused by a need of the probe retraction from a surface in every oscillation cycle. This circumstance limits a use of probes with stiffness below 0.3-0.5 N/m for imaging in oscillatory modes.

2. Engagement should be done most gently. In Nanoscope software under "Microscope" there is a window "Tapping Engage", the following choice of parameters looks most promising.

Engage delta set-point 0.01
Engage final delta set-point 0.00
Engage test threshold 100
Engage min set-point 50
TM engage gain 0.4
Sew tip Yes

Parameters below "sew tip" are not relevant after "sew tip" is "Yes".

While using soft cantilevers and the above engagement parameters one might experience a 'false" engagement especially when initial amplitude is rather small, say 0.5V. If this happens than further moving a probe down with a stepper motor in step-by-step approach might help to accomplish the engagement routine.

3. Light tapping conditions are preferable for scanning with sharp probes. In other words, free amplitude (or target amplitude) should be chosen rather small. Practically, 0.5V amplitude seems to be fine for imaging relatively smooth surfaces. After the engagement, set-point amplitude should be chosen as close as possible to the free amplitude. In other words, one should adjust set-point, scanning rate (slower - better) and gains (increase of Integral and Proportional gains can be done up to the point at which periodical oscillations appear). set-point amplitude is lowered from amplitude of free-oscillating probe One can achieve low-force imaging (light tapping conditions) by lowering set-point amplitude (from initial amplitude of a non-interacting probe) to the point that "trace" and "retrace" height profiles (in Scope mode) become close to each other and the phase of interacting probe is still not far from that of the free-oscillating probe.

Despite the fact that so-called true atomic resolution was achieved in the FM and AM imaging modes the value of these results is still questionable. The computer simulation of AM imaging of polydiacetylene crystal showed that real atomic-scale resolution could be achieved only when the AFM probe has an apex of the atomic size and imaging is performed in the low-force regime [11]. Otherwise, even the observation of the isolated atomic- or molecular-size defect does not mean that the molecular arrangement surrounding such defect is reproduced correctly.

Consequently, there is still a lot of room for further improvement of high-resolution imaging that requires the use of sharp probes and better quality instrumentation allowing low-force imaging using such probes.