Try to rotate the sample by 90°, if the streaks have the same orientation, then something with your instrument is wrong. Did you checked the quality of your tips? If you are investigating soft materials, then some of these soft structures may immediately stuck to your AFM tip even in tapping mode, so that they create artefacts parallel to the scanning direction - thats why I propose to rotate the sample!

Good luck, Dirk

I had a similar Problem, first try scanning slower, and than we should know your setup...i my case the cooler fan from one of the peripheral instruments produced vibration which resulted in similar line when scanning smaller than 1µmx1µm...

Is your AFM-Microscope alone on its Table?

It looks like some kind of noise. I would guess that either your acoustic or mechanical isolation isn't good enough. Try reducing your gains and see if that makes a difference.

I think there is some problem with cantilever or scanner so change ur cantilever and try again

The structure you obtain is quite repetitive. It thus looks like a artifact due to an additional oscillation of your cantilever. Try to change the scan rate and/or to adjust the P- and I- gain again.

It looks like some kind of noise

Have you tried with different scan size? What happend to streaks frequency? If you reduce the scan size, do the distance streak distance increase? In my opinion, it seems some artifact due, to some superposed frequency from outside. Illumination light? or electronic noise? Try also to reduce scan speed or increase tip distance from sample surface. Hope it helps. Erica

This looks like a periodic noise: it could depend on the bad instrument isolation from acoustic or mechanic noise. I had a similar problem with a fan making these stripes. Try to switch off every instrument close-by if you can. Good thing would be if you can isolate the table were the AFM is. I see this is a multimode AFM type: usually it is very stable...

Dirk's idea is good, although I reckon that finding the same spot after rotation is difficult. Try it with a calibration grid with some asymetrical features. The artifacts created by sticky tips are not as regular as your lines. I would say that it is some problem of the instrument: electronic noise, malfunctioning piezo,, defective mechanical insulation of the AFM head, etc...

I agree this could be due to the scan rate, 4 Hz is to fast for me. Try to repeat it at 1Hz

Now the spatial period of streaks is around 15 nm, does it change if you decrease the scanning velosity? Do the streaks gange they direction while you decrease scanning velosity? If the answer is "yes" try to improve mechanical isolation of your AFM.

This is a periodic noise arising from the power supply (probably). You can filter out

very easily in the fourier domain. Can use matlab image processing toolbox.

Mica is really flat and, if you look at your image, you have an almost perfectly periodically spaced ripple.

These are the images of the filtered and the noise image after FFT.

Sadly, it is too perfect to points any other thing than an electrical/optical noise in your system. My advise? Recheck all the conexions and try to find out possible noise sources.

This is noise image after the FFT

It looks like electical noise to me. By means of FFT analysis chek whether that noise has a peak at 50 hz. It s frequency should not be affected by changing scan size and/or speed. If this is the case, you have ground signal instability and probably you have to improve your ground loop quality.

It looks like some sort of background noise at about 4x the frequency of your scan frequency. (If you concentrate on a horizontal line, I think you get about 4 changes from light to dark, then back again). Is this consistent with something like electrical noise (50-60 Hz)?

If you decrease the scan rate slightly, then the diagonal pitch of the pattern should increase, and if this happens then as Erica Iacob points then the problem is with external noise. Good luck, Richard

It looks like your scan lines are interlaced. Checking your parameters you have selected the Retrace - try to change this to another option. It is a bit weird though that your streaks do not appear to follow the pixel rows in the image.

Every second line in the image seems empty - thus the cantilever is not in good contact with the material you are trying to image.

Sometimes you encounter DC drift related artifacts - but these follow the pixel rows in your image. If you encounter those, you can make SW routines that remove them quite efficiently. I can help if need

We had similar artefacts on a Multimode device, caused by the illumination unit of the video camera. They vanished, when we turned off the light (in our setup, there was a lamp housig next to the AFM and the light is coupled with a fibre into the video camera. We simply switched off the lamp housing during the AFM measurement)

Totally agree with Gabriel and Giovanni - this looks totally like electronic noise and that it is so perfectly filtered out by FFT as Gabriel demonstrated even further indicates that. The only other option I see would be some mechanical noise - check the frequency of the oszillation as Giovanni suggested, and see whether it is actually 50Hz or could be correlated to some other regular frequency around you in the lab. You could also try to shut down other devices in the lab and venting to see whether you can locate the source of the noise...

I feel , you reduce gain.

As mentioned before, the periodic structure strongly indicates external noise. I encountered similar structures while using a vacum pump to hold the sample. The vibrations from the pump were carried over to the sample by the air stream that should hold the sample.

It looks like the sensitivity is different between the forward and reverse traces. It is definitely some sort of noise, and it has to do with the settings you are using with the instrument. I'm not exactly sure what, as it has been a while since I've used an AFM. But is really looks like the forward trace is good, and you're getting nothing but noise on the reverse movement.

We had a similar experience of one of our AFMs. We did a lot of measurements, but could not find out the source. We even used different kinds of UPS machines with no effect. The most unusual thing was that there were days, where the features did not appear. And when strategically writing down all the possible differences between "the good and bad days", we found out that on the second flood in the same building (our AFM is placed in the basement) another researcher turned on a specially designed extrudor. The noise was the same as was mentioned somewhere above - in the range of 50Hz all the time. The feature did not disappear even with all possible tuning of hardware and software parameters. What we did is that we made an arrangement with the guy using the machine to prepare a schedule for measurements. But, mostly we had to measure at night. Then we moved to a new building, where all possible interferences were considered in the planning phase. Now we can measure the whole time.

I didn't go through all the comments you've got and maybe am repeating something that was already written, but maybe gave you a clue nevertheless.

Dear Krishnashish:

In addition to the recommendations of the other participants, perhaps you should check if the artifact that appears in your picture is because for periodic vibrations due to movement of an engine, next to the AFM. For example, the power supply of light that illuminate the surface of the sample, usually incorporating a fan to cool the interior of the lamp, try turning off the power supply, not only decrease the intensity, but turn it off entirely, so that the fan stops

If the cord or optical guide that carries the light to the camera is in contact with anti-vibration table, it is possible to produce this artifact.

Dear Krishnashish,

that was clearly electrical 50Hz noise from your grid circuit. Try to change the scan frequency and you will see how the period of noise changes. The point is that uses to build surface charges enhancing the noise. You can try to employ small mica samples glued with silver paint to the holder to reduce that. Otherwise, you must learn to work with the noise. As suggested by other colleges FFT can remove completely such periodical noise.

I hope this comment will be useful for you.

I see this with some of my images too when scanning at high speeds, I agree with Matteo and think a much lower scan rate should help.

I guess it is noise from some of the following.

1. May be from microscopy light

2. May be from head

3. May be from some kind of disturbance near by.

Or some cantilever problem

At a scan rate of 4 Hz, I do not think it is 60 Hz noise, rather gain is a little high. I am with others on this one, play with gain and scan rate a bit and should clean up.

Noise absolutely. Try to play as follows:

1. Check environmental vibration

2. Hold the mica stable on the stage

3. Change the gain number

4. Lower the scanning speed

The features that you see are almost certainly noise and from the banding of the noise it is possible that two main components are present. These components would appear to have a frequency of ca. 0.6Hz and 0.023 Hz, obviously these could be lower harmonic coupling to higher frequency noise. All of the suggestions you have had should be investigated to see if they change the dynamics of the noise.

If no vibration isolation system is present a simple and cheap solution which has been employed for many years is an inflated bicycle inner tube and rock slab resting on it. Then placing the microscope on top. This type of vibration isolation is particularly suitable to reducing low frequency noise. If nothing else it will change the vibrations in the system and enable you to rule out coupling from beneath the microscope. Another potential source of noise is acoustic pick up and for testing purposes a simple box lined with foam rubber placed around the scope can yield information. Lining the the box in aluminium foil can be used to test for EMF noise.

I would state the obvious as every one has said. "Its noise". and it might be coming from environment. You will need a better isolation. We had similar problems and we have seen 50Hz noise coming from even florescent light or fans in the amplifiers or your computer. Isolation is a must when working with sensitive instruments like AFM. Try using Constant Height Mode on some calibration sample for a smaller scan area. If the noise is still there, then its definitely coming from the environment.

If you dont see the noise in Constant height mode, then it might be your controller gains. Generally these systems are resonant systems and the manufacturers use simple PI controllers which bascially tries to low passes the resonant peaks. During raster scanning one of the lateral axis (say X) is excited by triangular signals and other with pseudo ramp. Now these triangualr signals have harmonics which gets amplified by the resonant peaks. If you increase the gains of controller these harmonics would be excited even further causing such vibrations in imaging.

Check the frequency of the noise. Many times has the exact same frequency of the main power line (60 or 50 Hz). If the case then try turning off unnecessary equipment.

Exchange your support and let's see what it gives us to determine if your cantilever do not need anything

In addition of noise that can be handle optimizing scan parameters: We observed some type of "noise" when the ligth of the optical microscop of the AFM was kept open

Accoustic vibration in the back ground (that can come from other instruments around) can contribute to "noise". In that case, better accoustic isolation would help.

Ask questions on certain scientific issues, I think is one of the best things and more clear objectives of the Research Gate platform.

The answers you get, in most of cases, offer interesting insights, and in of many cases are of great value, maybe for this, don´t forget , are given by people who take a part of their time to provide an answer.

In my opinion the least that is expected, when you ask a question and got answers, many answers, is that you show some interest and offer some feedback.

Thank you all very much for your answers to my question. It took me some time to try all that have been suggested in 40 answers. No matter how much I play with the gains or scan rate or scan angle or setpoint or drive amplitude or switch off possible electrical sources of noise, the rms amplitude of this noise never becomes less than 0.1 nm, even with FFT filtering. By decreasing the scan rate, the noise amplitude decreases and the spatial frequency increases to a point you won't notice any observable noise, but still there is noise which doesn't allow you to resolve the helical pitch of DNA. By increasing the scan rate, I am just amplifying the noise that's already inherent in the system.

Can anyone explain why there is beat formation in the noise or the FFT of the noise?

My plasmid DNA is about 0.4 - 0.8 nm high (it should be 1 - 1.5 nm). I used 50 mM MgCl2 to bind it to Mica. I used two cantilevers having stiffness of 0.5 N/m and 40 N/m; but no change. What is the reason for such a decrease in height?

Below you can see two images taken at perpendicular scan angles.

I think the problem is with the electric insulation/grounding of the system, which I can't get rid of currently, so my only option is to do post-processing.

I never expected to receive so many answers. I am so delighted and grateful to you people for finding some time out of your busy schedule.

Doing an FFT to get rid of the noise, removes any possibility of detecting the helical pitch of DNA. Can anyone help me applying a good filter to only remove the noise?

No change in spatial frequency of noise with scan size.

It is necessary to find the source. Have you tried using Constant Height mode on a Calibration Sample. If the oscillations persists in constant height mode then its definitely environment and not related to controller gains or resonances.

I haven't tried that yet as I do not have any calibration sample.

By the way, if I reduce the scanning frequency, the noise become almost unnoticeable, but not good enough to resolve the pitch of double helix.

You need to find the source there is no point speculating on the scan results. Which AFM are you using. Are you using the original setup of the AFM or you have customised some part of it. Are there any custom circuits you are using.

There is no customization with the instrument at all. I am using Multimode V from Vecco, bought in 2009.

Could anyone please comment on the lower harmonic coupling suggested by Dr. Christopher Muryn ?

He said: "The features that you see are almost certainly noise and from the banding of the noise it is possible that two main components are present. These components would appear to have a frequency of ca. 0.6Hz and 0.023 Hz, obviously these could be lower harmonic coupling to higher frequency noise."

In that AFM, I think there is just a small cover over the Z axis which serves as an isolation. Is there any other isolation provided with the kit.

The cantilever I used has a fundamental resonant frequency at around 116 kHz and a spring constant of 0.5 N/m. Do you think by using cantilevers of very high resonant frequencies, I will be able to overcome this problem?

Vibration isolation from Minus K Technology, an older version of this one: http://www.minusk.com/content/products/standard/bm4-anti-vibration-platforms-benchtops-isolators.html

What do you mean by "small cover over the Z axis"?

If you filter carefully your images by FFT you can save the helical pitch of the DNA because they have very different spacial periodicities. You could find issues if one of them were in the second or third harmonic of the filtered frequency, but as you have noticed, you can tune this by modifying your scan rate.

Anyway, as far as I can see, it seems to be a grounding/electrically coupled issue. We have had some problems with that and, sometimes, the noise just vanished after plugging all the (xPM+CPU+Display) electronics to the same plug hub and/or replacing the old CRT display with a new TFT. Maybe the CPU power supply could be messing all this up, but there is no way to be sure without a direct sight.

There is no way to connect an oscilloscope to the controller. However in the software you have something similar (Trace/Retrace vs. line). I have no idea what are you taking about exactly ... the tip is obviously not stuck to the surface.

The samples/line is 512 for all the images shown here. The main problem is that the DNA height is very less (about 0.4 nm) and when I go to scan sizes below 250 nm, the image gets completely washed out; all you see is noise.

Revise your electric power line. Switch of air conditioner if it works...

The noise pattern seems to indicate a low frequency noise source. My best guess is that this is vibrational noise, related to the vibration isolation setup or sampel fixing method used. Changing the cantilever likely won't change the result much. In order to test your setup for noise sources, you can scan 0nm on a flat, clean & solid-mounted sample (after optimizing the feedback settings: setpoint & gains). In a proper environment and when using proper vibration isolation, the RMS noise should be

My guess is building vibration. In our lab which is located on the 3rd floor, this is at approximately 10Hz and images look similar if you forgot to activate the vibration isolation. Have you set up your antivibration table correctly, or is it maybe broken?

Do you only see this noise when you are working on mica substrates? We had AFM noise that we only observed on mica, and that was attributed to the laser bouncing through the layers of mica. It just requires that the position of the laser on the cantilever is adjusted. While this may not correspond to the best sum, it does eliminate the noise.

Tanya, I've seen many noise patterns originated by laser reflexions on several types of substrates and, in my experience, all of them had really longer periodicities and are, at least, biaxial-like and not so uniform (It depends on the substrate) , resembling more like waves in the sea than ripples in a "Ruffles" chip.

Gabriel, Yes, fair enough - it has been a long time since I have looked at those images as we now have a new instrument with a different laser source.

The noise could be induced by the power line (just one option), in this case you can solve the problem using an UPS to "clean" the power line itself.

could be a ground loop, indroducing some oscillations at the data acquisition. A ground free transformer to separate the controller from the grid could solve the problem.

Did you use the commercial system or did you add some custrom made electronics in the loop for signal processing? Osciallations in custrom made op-amp circuits or a slight frequency mismatch of sampled systems can also cause such effects.

It seems to be feedback closed loop generation. If the initial tip oscillation amplitude is too high, and sample is very flat (like something on mica substrate), the signal induced by the Van-der-Waals force from changing topography is too low in comparison with oscillation amplitude, there should be large force between sample and tip to get enough change to resonant frequency, AFM become very sensitive to all sources of noise (electrical, mechanical, acoustic), FB gain become too high for such SNR and feedback loop began to work as oscillator with positive feedback. It can be clearly checked by computing 2D Fourier spectrum (or using Fourier filtering), there should be vertical lines, perpendicular to fast scanning axis, in this case.

Are these lines always parallel to the horizontal axis ?

It is harmonic generation with period(s) defined by feedback loop delay and usually it is perpendicular to fast scanning axis (if you clear the highest harmonic from image above, you can see slight vertical lines on Fourier spectrum). Beside this generation, the ill-tuned scanning system can catch another sources of noise (50/60 Hz harmonics from electrical noise, around a 1-3 kHz in acoustic noise from ventilation and electrical from inverter and so on). I have seen up to 5 different noise frequencies in very bad tuned scan in my practice. As noise is sampled with regular interval, it frequency can move to the low freqs. region as difference between sampling and noise freqs. (an so it will appear along slow scanning axis).

It is very likely electronic noise was picked up during the scan due to electrostatic charge. A simple check is to slightly change the tip frequency to see whether the noise attenuates.

If it is the case, a simple solution is that you can lower the laser sum and I gain to reduce interaction. But this affects image quality. Sometimes, I found pushing the tip hard to surface might also help. To uprooting this problem, it necessitates ground to sample or sample holder. I used carbon tap before. If none of these help, you need to contact the company to check the AFM. There might be broken components.

If you have another same type AFM in your premise, I suggest try to swap the controller to take a look on the effect. I did have an experience where the 60Hz noise is still appear even at 1nm scan size which it indicate the noise is coming from the electronic parts. I believe you have tried many settings to troubleshoot this problem. If you are not confident enough to swap the controller, try to swap the smaller parts first (eg: probe holder, scanner head) but please ensure that the scanner head file should be copied into your system before you use another scanner head.