Asylum Research

Galleries | Cypher Fast Scanning


Celgard, the Bruker Challenge Sample (tapping mode)

In recent posts on LinkedIn and on the SPM Digest (see footnote below) Bruker suggested that Celgard® should be used as a fast scanning standard.

We at Asylum Research agree.

Celgard is an interesting material that is used as a porous battery separator membrane and as a filtration membrane in medical applications. It is made by stretching isotactic polypropylene. The topography has sections with drawn fibrils (which form the pores) and also sections with undrawn crystalline lamella. The fibrils form something like a 1D grating and are challenging to track. Because the fibrils are freely suspended, tapping too hard will displace them. Lighter tapping gives rise to a slower cantilever response, so for this sample having a fast cantilever as well as a fast Z actuator is important.

In repeating (and beating) the Bruker results, we matched as closely as possible the scanning conditions in their movie. Because of small differences in how we do the turn-around at the edge of the image (sometimes called rounding), our linear velocity ended up being slightly faster than Bruker’s (23.6µm/s vs their 22.5µm/s) while our scan rate ended up being slightly smaller (9.8Hz versus their 10.5Hz). For monitoring the tracking of the Z feedback loop, the linear velocity is what is important.

In assessing how well a fast scanning AFM is performing, looking at the alignment of the trace and retrace of individual scan lines is a great diagnostic. This gives you information on how well both the XY loops and Z loops are tracking. Luckily, most manufacturers display the trace and retrace data from the current scan line below the real-time image.

The following image shows the trace and retrace from a frame from the Bruker movie (see link below) versus a frame from the Cypher™ movie (below).  Even at a faster linear speed, the Cypher is tracking the Celgard surface much better than the Bruker FastScan. In particular, there is good agreement between trace and retrace even in the very steep 40nm "canyon" between two fibrils near the center of the image (see red circles for comparison). The Bruker trace and retrace show significant differences in both shape and lateral offset in all the canyons indicating insufficient Z bandwidth. This trace-retrace diagnostic shows that Cypher is returning a much higher fidelity representation of the surface shape under the same scan conditions.


Watch both movies and judge for yourself:

Our Cypher movie is here:

(Play Cypher Celgard Movie)

Click here for Bruker Celgard Movie: YouTube


Footnote (from

I would be interested to hear the community’s thoughts on the quality and speed of these images, particularly because Celgard is such a challenging sample. Again, unlike hard flat samples (which several movies have been put forward), Celgard really challenges the z-scanner speed, by making it track large excursions (top to bottom of features) while maintaining refined force control (seeing the “blocks” of the mesh). In an earlier thread on Linkedin it was suggested that Celgard be an imaging standard, using as the quality reference: Fig. 13.25 and 13.29 in the book "Surface Analysis with STM and AFM" Weinheim VCH 1995. I have excerpted those figures from the book and posted them here for review. Scanning >10x the typical could be the benchmark for claiming a high speed AFM. I believe all vendors also have access to Celgard, but if anyone out there thinks they can go faster, but does not have it, let me know and I will send it you. – Steve Minne, Bruker AFM, on the SPM Digest, May 2011

Oxford Instruments Asylum Research, Inc. • 6310 Hollister Ave. • Santa Barbara, CA  93117 •