Hello Yong,

a bit of drift is quite inevitable with microscopes, and some correction methods are available to compensate for it, both during the experiments, or offline on the images data. With confocal laser scanning techniques it is usually enough to acquire 3d stacks on a field of view slightly larger than necessary, and compensate after the experiment by re-aligning the images. You can find a nice and quite simple method here:

http://onlinelibrary.wiley.com/doi/10.1196/annals.1342.029/full

That said, 20 microns in six hours sounds like a lot, if i were you i would try to contact Olympus, or at least find the datasheet/manual of your stage, and look for an expected stability specification.

On most manual scopes if you grip the right coarse knob and the left coarse knob and rotate them in opposite directions it will tighten the z-direction axis a bit.

Otherwise, maybe you could tape the coarse knob in place?

If the whole microscope it is not enclose within a plexiglass cabin and only the stage is warmed you can try to minimize it by using an objetive heater

Dear Paolo Pozzi,

Thanks for your kind answer. I will check that. 

Dear Jordan Robin Yaron

I didn't hear that kind of functions exist in my microscope yet. 

I will check it asap. Thanks!

Dear Victor Campa

My system was shielded with blackout curtain(one side) and some formboard(kind of styrofoam, three-side) and Object lens heater was set to same temperature with stage. 

Thanks!

Yong, the reason i asked about the enclosure it is because ideally the whole system, and not only the stage, shoulbe be a 37ºC. We even keep the inmersion oil at 37ºC because we found out that at RT was causing us focus drift. We also turn on the microscope with enough time to equilibrate at 37ºC and let cells to settle for at least 30min before launching the time lapse.

Other additional concent are photodamage (laser scanning it is very hard on cells). Conventional epifluorescence + deconvolution it is usually better for the cells. And if you really DO need confocal, it is better to use a spining disk (is avaliable).

Finally, for >6h of time lapse I would consider a CO2 control system too. HEPES-buffered media might not be enough.

Dear Victor Campa

Thanks for your kind answer.  what "RT" means exactly? thermal circulation of immersion oil between object lens and specimen?

And I am using two-photon excitation that known to cause less damage than that of  single photon excitation. But I agree with you that laser itself is harmful than spinning disk.  Can I ask to you whick kind of deconvolution method did you use for epifluorescence? I think epifluorescence couldn't make good resolution.

Regards,

Hi again Yong,

RT just means room temperature. If we use the oil at room temperature and the microscope was at 37ºC there was a huge focus dirft. Right now we only have a licence for deconvolution for Leica´s software LAS-AF but some researchers that have worked with Imaris say that it is pretty goog. Hopefully we´ll get one license in the near future.

Choosing one sytem or another depends on the kind of experiment, but in general I believe that two-photon suits more exp with animals or pieces of tissue were you want to prenetrate depth and spining disk or just epi+deconv works better for cells monolayers. Sometimes it a matter of trying. Anyway we do not have a two-photon here and I can only choose between single photon LSM or epifluorescence for setting time lapse. I´d say that >90% of the times we go for epi.

Dear Yong,

I agree with the above comments.

1. Equilibration of all equipment and reagents to the temperature required for your experiment is critical. For live cell imaging: 37ºC + 5% CO2 for at least 1hr prior to commencement of experiment.

2. During your imaging collection expand the region you will capture in z. This will help compensate for some movement in your experiment.

3. There are a number of post collection methods to correct for drift using freely available software. I have published a method using the ImageJ (Fiji) plug-in called: Correct 3D Drift. This protocol can compensate for sample translations across four-dimensional confocal datasets.  A link to this paper is provide

Manuscript PMID: 24747942

Hope this helps. Good Luck

http://www.ncbi.nlm.nih.gov/pubmed/?term=24747942