Dear Tanu,

this topic is indeed not so easy. You could argue that with large enough external magnetic fields, you should be able to switch the complete AFM so that the EB vanishes ... and following this might lead to the idea that the EB is always an effect of minor loops.

However, from a pragmatic point of view, we should restrict ourselves to reasonable external magnetic fields. Can you rotate the sample in the cryostat? If so, you can simply cool it through the Néel temperature in an external magnetic field, measure a hysteresis loop, rotate the sample by 5 or 10 deg, measure a hysteresis loop again, and so on. After 360° measurements, you change rotation direction and measure "back" to 0°. Then you compare the EB values for both rotational directions. If they are identical, you don't have a problem with minor loops; if they differ in some areas, then the maximum external field was too small there.

I had some samples with very pronounced fourfold hard axes for which I always got an asymmetry around the hard axes which was simply due to (invisible) minor loops, while around the easy axes everything was fine. We made some simulations about this topic in the attached publication.

Best regards,

  Andrea

Article Influence of fourfold anisotropy form on hysteresis loop sha...

Interesting answer. How do you deal with training effects in this context?

I measured Fe/FeF2 and Fe/MnF2 samples and found pronounced hard axes especially in the latter. On the other and, there was nearly no training effect, quite opposite to Co/CoO and similar systems. However, in Fe/MnF2, even with a training effect it would have been easy to recognize whether the hard axes are shifted by ~ 5 deg - they were quite "spiky", not "rounded" as in Co/CoO and also Fe/FeF2.

So, experimentally, the first thing would be to convince oneself that training effects are irrelevant and next proceed to the variation in angles and the more quantitative evaluation of what is found then? I think I start to get it...

Sounds good - thanks for reminding us of the training effect; I was so focused on the system Fe/MnF2 where I had the minor loop problem too that I completely forgot to think about this issue.

Thank you for the discussions...

actually my sample is polycrystalline antiferromagnet, but I have small hysteresis loop opening with the strong antiferromagnetic straight line MH, and quiet surprisingly it is showing exchange bias, so first I wanted to confirm that the observed exchange bias is not due to minor loop effect.. I have tried testing training effect as well, and it is showing training effect..

about the suggested rotation of the sample,it may not help because the sample is not oriented.. it is in bulk form... but I will definitely try this technique for my thin film oriented samples..

Dear Tanu,

could you show us one of your Hysteresis loops?

If the sample is completely polycrystalline, a possible EB should have the usually claimed cosine form. This means that you can field cool at 0 deg, measure a loop (probably with EB), rotate at low temperatures to 90 deg (-> EB should vanish), -180 deg (EB should have the opposite sign) and 270 deg (-> EB should vanish again).

Seeing a training effect, by the way, is not singular to EB systems but would also happen in minor loops.

Best regards,

  Andrea

Dear Sir,

we are doing measurement in SQUID magnetometer, so I can not rotate the holder containing the sample, but what I can do taking out the sample and mount it opposite way but in that case 90deg precision would not be there... but will that help?

here i have attached my Exchange bias MH loop.. the green one is ZFC MH, whereas the red one is FC MH

Dear Tanu,

thanks for the picture! But then everything is clear - you can only cool in a negative field again and see whether the EB is positive then. And of course you can do the same in other sample orientations. But as long as you see such a large difference between FC and ZFC, there must be an EB.

And additionally, if you have a SQUID, you can measure the complete temperature dependent magnetization for FH after ZFC and afterwards FC and see whether it looks like a pure (maybe diluted) AFM or whether there are any hints on a ferro-/ferrimagnetic fraction.

Best regards,

  Andrea

I have done negative FC MH loop, it is shifting vertically downward and to the positive side of magnetic field axis... just the opposite way of positive FC MH,

thank you for kind reply .. definitely this is useful for me

So actually you do have an EB. Interesting - any idea where it may come from? Which AFM do you work with?

I work on rare earth manganite system, I am still looking for the origin of exchange bias.. probably XMCD measurement can help

generally spin glass system or phase seperated manganite system can give rise EB

The interacctions in the interface are important. Due to if the spins are compensated in the interface,  there are a  enhancement of the coercitive field, in case contrary the coercitive field decrease, due to interaction in the interfase. this is one of the reasons by the that not saturate you systems

Are we talking about, as your OP more or less suggests about ferro- or ferrimagnetic particles in an AFM matrix? Yeni was right to emphasize the interface. So, what IS the interface in your case?

Dear Tanu,

according to your graph in the fields above ~800 Oe, the hysteresis loop collapsed and the total magnetic moment Mtotal is a linear function of the magnetic field H. This indicates that in the fields above ~800 Oe, the magnetization of the ferromagnetic (FM) phase has reached the saturation, while the magnetization of the antiferromagnetic phase continues to increase. Since the magnetic moment of the antiferromagnetic (AFM) phase depends linearly on the field, you can separate the magnetic contribution of the FM phase to the total magnetization, by subtracting the AFM contribution from it:

        MFM = Mtotal - MAFM = Mtotal - (dM/dH) * H

(where MFM is magnetic moment of ferromagnetic phase, MAFM is magnetic moment of AFM phase, Mtotal (or just M) is your experimental moment, and dM/dH should be calculated in your case in fields more (!) than ~800 Oe, where the hysteresis loop is collapsed).

After this procedure the coercive force and, therefore, exchange bias field will be changed. And may be your loop (wich were received after cooling in field 1.5T) will be shifted only on the field axis.

More info you can read in the article:

Article Exchange bias phenomenon in (Nd1-xYx)2/3Ca1/3MnO3 (x = 0; 0....

I don't think that the conclusion in the previous answer is necessarily correct (I am not claiming it's incorrect, just voice my opinion that there might still be more than one scenario compatible with the data). If the "up" and "down" branches join above some field, then this "only" indicates that beyond that field there is a region where the magnetization change is not caused by irreversible switching processes. It could also (partially) be due to a rotation of magnetization in the ferromagnet, couldn't it?

Dear Yeni,

I am not comfortable with interface effect of the compensated/uncompensated spins. so if you explain it  or send me some link where the correlation of the coercivity is explained that would be very nice. you told that coercivity will increase if the spins are compensated, but as per my knowledge spins in a ferromagnet are uncompensated...

If I am wrong please make me correct... Thank you.