If the substance exhibits a high degree of symmetry in the crystal and much lower in amorphous form, certain signal may appear in the amorohous form which are not observed in the crystal. Selecton rules predict nonzero IR absorpption intensity only in vibration belonging to certain point groups (which transform in the sam way as the x, y and z coordinates). You should tell more about the substance, provide the spectra as well as the the XRD observations - otherwise the question is too general.

Substance is an organic heterocycle with an oxo and an ethyl attached to the same ring. My colleague send me the view obtained from XRD experiment, which is a monomer. In another pic he sent me the view of 3D supramolecular network generated by inter-molecular C-H...N, C-H...O and N-H...O type hydrogen bonds. After I've examined the cif file in Mercury, I've decided that the repeating unit in the crystal structure is a central-symmetric dimer. I've attached both IR spectra, upper one is for powder sample and lower one is for crystal sample.

IR spectra have small variations due to particle size. In order to make sure there has not been any interconversion within the sample, ID confirmation with another independent method would be useful (NMR, perhaps MS)

Looking at the actual two spectra I am pretty sure that there is a chemical difference between the two samples. The strong OH (or NH?) signal and the appearance of strong peaks above 1500 1/cm clearly indicate the presence of new compound in the second spectrum. The differences are far larger one would expect for size or symmetry effects.

I guess so. At first I was suspected from polymorphism, but it doesn't look possible.

It would also be helpful to also try Raman spectroscopy as the spectra is usually simpler than that of IR although the two techniques are complementary. This may also remove the ambiguity associated with the broad bands that are observed here and indeed lead one to conclude that these are different compounds.

There can be an orientation effect for the single crystals, depending on the space group and the orientation relative to the polarisation of the beam. In ATR the evanescent field that "touches" the sample has a preferred, though not total polarisation. You might try rotating the sample on the ATR crystal.

Hello, crystalline samples might be a nuisance in ATR-based FT-IR analysis, because they reflect the beam and give false results. It is advised to get rid of crystals by crushing them into powders or consider the good-old KBr pellet approach. Best wishes on your research.

An alternative answer would be the presence of another compound in the non-crystalline sample and the crystalline sample being the pure one in your case. You will decide if it is correct or not. Greetings.

Try to perform a measurement on a powder withouyt water ..... Your powder spectrum is strongly dominated by OH lattice vibrations .... So I think you obtained a new compound due to water contribution ....

Also under clean conditions the IR spectrum of a powder may differ from that of a crystalline absorption spectrum if the vibrational oscillators are strong. Then, resonant scattering by the powder grains takes place and spectral features can be dominated by the scattering cross section and not by the absorption cross section. For spherical particles this behavior is described by the Mie theory.

The differences between the two spectra suggest that the lower one is probably a hydrate given the bands at 3460 and near 1640cm-1. (The latter is water not simply OH) However the presence of a strong band near 1000cm-1 in the powder with nothing similar in the crystal is puzzling if water of crystallisation is the only difference. If this is a single crystal then orientation can have dramatic effects on ATR spectra. The suggestion of rotating the sample is good but you should rotate about axes in the plane of the surface of the ATR element not simply about an axisnormal to it.

I agree completely with Richard Spragg. The lower spectrum (powder?) has picked up water, possibly both specific hydration and surface water (on the other hand, if the spectra are reversed, this may mean that the crystal has lost water on being ground up). In the first case, eliminating that by careful vacuum drying and anhydrous sample handling will probably eliminate most of the difference (presuming the samples are otherwise chemically equivalent). The polarization effects, especially with the single crystal (and possibly even artifacts due to poor contact and/or polarization with the crystal and the ATR element) might account for some of the smaller differences.

The lower spectrum is crystal sample. Depending on the DFT calculations I may say that the band at 3461 is v(NH).

Dear Serdar, as crystal do you mean bulk? If so it is very straightforward that powder and bulk spectra are very different.

You can see an example for calcite here:

Bulk http://speclib.jpl.nasa.gov/graphs/speclib_1357637278_2013_01_08_plt.png

Fine powder http://speclib.jpl.nasa.gov/graphs/speclib_1357637306_2013_01_08_plt.png

There will be difference in the intensity of the peaks with respect to the phase of sample. There should not be any difference in the frequency. Vibrational modes are within the molecule, so basically an isolated molecule is necessary for recording the vibrational modes which falls in IR range. So, the gas phase is ideal one. If it is liquid or solid there is intermolecular interation which will have some influence mostly on the intensity only not on the bands. Even if there is some thing on the frequency, it won't as you think to be.

If you are dealing with some organic compounds, the difference could be due to its salt or base form. Another reason is what has been pointed out in the above thread, crystals will give problems of poor optical contact. But this shall decrease or increase the general absorbance intensity and not to the extent that you lost few peaks. It is advised that you state the name of the target compound and tell us in detail how you analyzed the sample.

Yet again people are being asked to comment on an interpretation with only partial information. Attempting to interpret a spectrum without a structure can be extremely dangerous especially when the spectra have undergone data manipulation as these clearly have. A background "flat" has been applied which is causing peak distortion and relative peak height errors. Do not do it. It lokks pretty but loses information. Having said that I agree with the two Richards and Gyorgy. The band at aprox 3400 cm-1 is, from the shape, probably OH and the strong bands in the 1490 cm-1 region suggest more than simple amorphous/crystalline changes. The broad bands around 300 cm-1 in both spectra show strong hydrogen bonding such as carboxilc acid OH or amine salt NH. From the shape I suspect the NH. The pattern changes between spectra so there is clearly some change in the hydrogen bonding system.

XRD is able to measure crystalline (oriented) as well as amorphous (dissordered) phases in a substance. FTIR or IR is able to measure assymmetric vibrational modes and hence if your sample is crystalline or highly oriented you should experience or observe a simplification of your vibrational spectra. In general, lowering of the symmetry in the system will produce more observable vibrational modes in IR. Hope this help to explain your observations.

I already voted up Richard Palmer's response, but I'm adding another endorsement here based on knowing him personally as a student of his at Duke in the 1990's.

Thanks for all your valuable opinion and suggestions. Hope I will solve this issue when I will be free. Nowadays I should take care of the final exams of undergrads.

Good luck!

I came across this "old" question and I think one of the most important points of the original question has not been touched upon. While scattering has a minor influence in the IR, anisotropy is strongly dominating. As a result, it makes a big difference if in a randomly oriented material (powder, pressed or sintered pellet) the individual crystallites are small or large compared to wavelength.

In the first case random orientation leads to a reduction of the dielectric tensor to a scalar (but the intuitive solution of taking the sum of the principal components of the tensor and divide it by three is wrong, things are more complicated).

In the second case, the large crystallites, light is doubly refracted in each crystallite resulting in blue-shifted bands, altered band shapes and a depolarization of light. The most famous effect related to this is a band existing in coarse-grained polycrystalline hematite that is absent in the fine-grained form and in the principal spectra of single crystal hematite.

For more details see the link below:

Article Optical isotropy in polycrystalline Ba_ {2} TiSi_ {2} O_ {8}...