It is easier to assess this graphically rather than by inspection of the numbers. I have attached an image showing the two PDF patterns for comparison. I would say that there are enough differences between the patterns not to cause any problems in making an identification one way or the other. However this does not preclude that your material might contain something other than these two patterns.

One issue however might be that if your material is nano sized, the peaks may be too broad to achieve the resolution required. In that case as has been mentioned, full pattern profile fitting may be the best way forward.

Can you show us your data?

It may be possible to deduce the presence of overlapping peaks by using appropriate software to fit peaks to the XRD pattern in regions where you have a broad peak that looks like it might be a combination of peaks rather than a single peak. I haven't used XRD, but I've done this with XPS spectra, which can have similar problems with overlapping peaks.

The answers to this question may be useful:

https://www.researchgate.net/post/Can_I_fit_XRD_peaks_with_a_Gaussian_function_a_Lorentzian_Function_or_a_mixture_of_the_two_How_can_I_find_the_percentage_of_Crystallinity

Near 2theta angle (if Cu k-alpha is used) of 40 degree you have one peak in both the samples. In alloy In3Sn it appears after 40 degree (002)  and in pure metal the peak appearing at 39 degree is of 110. It is possible to analyze (qualitatively) the different phases present in the sample by doing simple profile fitting. 

I think FT-IR spectroscopy can be helpful in order to find out that, which one of In or In3Sn phases is formed. If you see the peaks that are assigned to In-Sn bonding, it means that you have In3Sn phase in your sample.

Thank you for the answers! But some of my specimens did not have any peak at either 41.36 degree (In3Sn) or 39.16 degree (In), so I am not sure if I would be able to distinguish those. Also, I know that XRD peak will shift due to change of strain. I am not sure if the 2 degree difference between 41.36 and 39.16 could cause by the shift.....

It is easier to assess this graphically rather than by inspection of the numbers. I have attached an image showing the two PDF patterns for comparison. I would say that there are enough differences between the patterns not to cause any problems in making an identification one way or the other. However this does not preclude that your material might contain something other than these two patterns.

One issue however might be that if your material is nano sized, the peaks may be too broad to achieve the resolution required. In that case as has been mentioned, full pattern profile fitting may be the best way forward.

Can you show us your data?

Hello Dr. Slipper, your answer is very helpful. The raw data file is attached. Compare to the standard patterns, the sample contains Sn, InSn4, and maybe In or In3Sn. Since there are two peaks at 32.9 and 36.12 degree, I am not sure if I can say they belong to In instead of In3Sn. 

The spreadsheet appears to be empty

In general it is possible but as you described your materials are supposable nano sized which leads to a broadening of the reflection observed in XRD measurements, hence the reflections will overlap (strongly) this might render it impossible to determine the relative amounts of both phases but the identification shouldn't pose a problem.

Sorry, I updated the post. 

Hello Yang Shu,

thank you for the data. Fortunately your crystallite size is not that small for a nano material, so the peaks are good and sharp. I have made a qualitative comparison of all the patterns in PDF-2  and have presented an image below with my findings. 

One comment. You must gather  better data ( I sound like a stuck record) The signal is noisy and the peak shapes are poor. If you count for longer you will get better data. Please everybody that reads this that is doing XRD obtain the best data you can, don't settle for quick noisy signals, it makes your job so much harder!! Rant over.

Anyway - the pattern for In0.25 Sn0.8 (00-048-1547) fits better than In3Sn (07-0345). The intensity for In (03-065-9292) fits slightly better than In (05-0642). There is also pure Sn (03-065-7657). There are other peaks as yet unidentified, but if you can regather higher quality data then maybe they can be identified. I hope this helps

Hello Dr. Slipper, your suggestion is very helpful. So for nanoparticles, do you have any suggestions on how fast I should scan? 

The phases that you identified from the analysis are the phases I expected. Thank you!

Yang Shu, scan speed is a matter of experimentation, there are no hard and fast rules. It depends a lot on what detector you have and a lot on what type of sample is being examined. It is a trade between the time spent on the data gathering and the quality of the data. Any X-ray counting system is subject to random events, so by its nature, the longer you count at a given point the better the signal to noise ratio. If you think there might be a peak in the background but can't quite tell if it is noise or a signal, then you need to increase the count time. If the signal grows then asses the peak shape - if it is ragged, you need to increase the count time. The aim is to have a really good signal/noise with well formed peak shapes; then any data analysis will be easier. Sometimes, the sample is so poor at diffracting that it is quite difficult to achieve, but try anyway. Do not get trapped into using one set of diffractometer parameters for all samples. I always run a rapid reconnaissance scan first, then assess the quality (Decisions to be made - have I got all peaks?=start,stop; enough data points in a peak?=step size; signal/noise?=speed; peak breadth?=slit size). Then I will adjust the parameters and go for data collection.

I have attached a good presentation by Bruker AXS on XRD basics, there is some treatment of data collection strategy here. it may be helpful.

Excellent summary Ian! Thanks for sharing the Bruker manual!

Seldom do XRD data sets display such parameters. Most diffractograms displayed are not even normalized along the "theta" axis to Sin(theta)/Lambda. It makes comparison and analyses difficult.

These are precisely the reasons we developed and used the real time 2D imaging system. It is a whole different paradigm with such a device! An example with a corundum standard powder at a scan rate of 10 DPM (degrees per minute) is attached below. From 20 to 130 degrees in 11 minutes. The sample to detector distance was 110mm. At this distance the 28um pixel size would subtend 51.6 Arc Sec at the rocking axis (R.A) or diffractometer axis.

https://www.flickr.com/photos/85210325@N04/26541827325/in/dateposted-public/

https://www.linkedin.com/pulse/powder-characterization-waxs-wide-angle-x-ray-preferred-ravi-ananth?trk=mp-reader-card

"Can XRD distinguish two crystals with similar peak locations?"

In principle, absolutely! Provided you have the necessary resolution in reciprocal space. You would have to consider the whole pattern to evaluate this aspect. A few peaks may match but not all related peaks will, unless they are the same phase.

Since you were talking about time...please consider that statistically the signal to noise ratio scales with the square root of the time, i.e., it doubles if you measure four times longer. A ten times longer scan roughly only improves your signal to noise ratio by a factor of three. For short scan times is might be remarkable since 10min or 100 min scan is no problem, but 3h or 30h it becomes. Then other opportunities might become interesting like an increase of the x-ray flux using bigger aperture slits or increase current or acceleration voltage, or use special optics. All of them usually have specific disadvantages which is pointed out by Ian when he mentioned there is no general rule. Time is certainly the most simple but also least efficient way. 

Exactly! If you have ICDD database ou can distinguish these peaks or any different phases.

Dear Yang Shu 

It'll be better if you use energy dispersive X-ray spectra (EDX)‎ to distinguish between

two crystals with similar peak locations.

@Gert, agreed. One of the easiest ways to increase the counts is open up the slits.

@Essam, EDX will not help in this case as this is an intimate mixture of nanoparticles, not discreet crystals.

Ian, Thank you very much for your correction

Hello Dr. Slipper, I ran the sample with lower scan rate, the noisy is much lower, and now it will be more clear to determine the peaks are In or In3Sn. Thank you! 

Yang Shu! How far are you from MIT in Cambridge, Ma.?

Want to see a demo of 2D real time powder diffraction? Real time Debye-Scherrer Camera!

I'll attach a sample YouTube video for the Corundum standard powder using a Bruker D8 in the Bragg-Brentano configuration as well as other data sets. So hypothetically, in your case if you can identify the peaks of interest, then you may be able to conduct dynamic experiments as the crystallization is occurring.

In 2D you'll be able to identify any presence of preferred orientation besides multiple phase development in the sample in real time.

https://www.linkedin.com/pulse/powder-characterization-waxs-wide-angle-x-ray-preferred-ravi-ananth?trk=mp-reader-card

XRD patern for samples having the same peak can be solved by the difference in the d spacing values for peaks and the internist of each peaks all that reflect to the planes of  the two samples

In addition, it can be solved by applying Rietveld Refinement analysis on the XRD data.

No, you will lose the peak positions for the both crystals, where you got only one broad (not sharp) peak for the both crystals, the result which is not correct.