In my experience, most of weird I-V curves are due to bad contacts, too high speed scanning, and temperature changes during the measurements. In rear cases, with complex cell structures, two cells with different Uoc are effectively created in series or in parallel, but these cases normally lead to small FF. Wet contacts, thin metal layer oxidation (Ag, Cu), organics degradation in air and/or under illumination.

I'd recommend to make scanning in two directions in all measurements. If they look the same, one scan can be discarded. If they aren't, you immediately see the problem, at least its existence. Usually the double scan doesn't increase time of the whole experiment much, as installation of the sample, making contacts etc takes the same time for single and double scans, longer than the scan. 

Dear Eidson

I agree with Vladimir answer about series resistance and value of fill factor. I say the following observations:

1- FF has high value and if you observe the value of Isc grater than Imax (i think that not right) where fill factor if we say excellent device the Imax was near from Isc and in the best cases was equal to it and this mean FF value equal 100%, therefore i think the fill factor must less than founded.

2- Why value of series resistance not found

3- value of Voc is very good and this mean the structure of device is good and not have much pinholes.

4- What is the value of incident intensity, are you sure take it 100 mW/cm2.

5- Are you sure the area is 0.04 cm2, this is very small for sample under test, therefore this effect on the efficiency and addition to the incident intensity value.

Finally, i hope i can help you, and this informations from my experience.

With my best wishes

Thank you very much for your information, I will take care of what you said, and I will clean all the contacts as well to be sure that is/is not the problem. Regarding to the two direction scanning, I not sure but I think it was observed when reverse mode. being that the case what can be the conclusion? Regarding to the other questions sure the irradiance is 100 and the area of the cell is 0.04 cm2. I usually prepare devices that have 4 cells on it, which all of them present the same area and we test them individually, so that is why the are is to small, I was thinking to connect them in series and then measure all in once, with an are of 0.2.  

Dear edison 

It is true that the characteristic is bizzare, I agree with the remark concerning the FF which is too large.  For the calculation of the efficiency, what the value of  the intensity power because I think the effi is as great.

very interesting. you should show a dark i-v.

This is the same question?

https://www.researchgate.net/post/I_am_observing_a_strange_behavior_in_my_I-V_curves_Can_anyone_tell_me_why_is_that?tpr_view=ZJEhBC3hKO1mhbEoG10d3ha9Eprg9z2FyGGB_3

Hi Vladimir, yes Olivia is working with me. And both of us are trying to find some answers. Thank you. So far we have done the measurement in two different solar simulators, and the results are similar, and also the film factor is more that 90. It is hard to believe it but we don't know what is wrong yet. Our efficiency is around 18%, but in some cases the shape of the curve is weird, that is not always happening but it frequently.  

Is it a single-junction cell or a tandem structure? For tandems, a current matching problem can easily result in such an I--V curve. For organic cells, nonlinearities in the series resistance may occur which could lead to such an I--V curve.

It is not a tandem structure,  the architecture is ITO/PEDOT-PSS/Perovskite/Fullerene/Al. But is such behavior is observed in tandem, what can be the explanation? Thank you in advance. 

Did you check https://www.researchgate.net/publication/261044878 already? Have a close look at Fig. 1: For the lowest scan rate, 0.011 V/s, and going from forward bias (FB) to short circuit (SC), there is a reduction in external current density for smaller voltages. (I have no idea whether this is relevant for you, I just try to give hints, since this is not my field of expertise.)

Article Anomalous Hysteresis in Perovskite Solar Cells

Thank you very much. 

I guess this graph is result of simulation for certain type of PV-modell with strong Temperature-Current Feed Back.

I think you must reduce the scanning rate (increase sweep points) . If your device based on liquid electrolyte it might be excess amount of electrolyte 

Since you said it is a perovskite, this behaviour is most probably related to some hysteresis effects. Try to drastically reduce the scanning speed and maybe also try to revert the sweep direction and see what happens.

Thanks for your support

What about of the effect of shant and series resistance on your simulation

If you change sweep-direction, you can find I-V isn't closed overlap for emerging PV. This is called " Capacitance effect " and "  Hysteresis effect" for Emerging PV, Ex. PSC (perovskite solar cell), DSSC, OPV, ... 

The solution, you will be find from

(1) "Towards stable and commercially available perovskite solar cells", Nature Energy 2016, 16152 (DOI: 10.1038/nenergy.2016.152)

(2) https://www.researchgate.net/project/Performance-rating-measurement-of-emerging-PV-for-Outdoor-AM15G-and-Indoor-lighting

I hope that can help to you !

best regards,

Mickey

Thank you very much to all of you that in a different way were providing me with an explanation or suggestion for this issue. After some time we were able to solve the problem and it was solved when the entire device was fabricated inside the globe box. Previously we put only the perovskite and the fullerene inside, then we took them out and transfer them to the evaporation chamber in air conditions, now we transfer them through a tunnel and no air is present. Still I don't have the explanation for this, but we haven't see that weird shape anymore. Thank you again.