ATR is one accessory to measure FTIR spectra. There are three types of accessories namely Transmission, ATR and specular reflectance.
FTIR-ATR is ideal for strongly absorbing or thick samples which often produce intense peaks when measured by FTIR transmission. FTIR-ATR will have penetration depth around 200 nm depending on your ATR crystal material.
so for power samples I recommend FTIR transmission.
ATR is one accessory to measure FTIR spectra. There are three types of accessories namely Transmission, ATR and specular reflectance.
FTIR-ATR is ideal for strongly absorbing or thick samples which often produce intense peaks when measured by FTIR transmission. FTIR-ATR will have penetration depth around 200 nm depending on your ATR crystal material.
so for power samples I recommend FTIR transmission.
You don't specify what you want to measure in the powder, but if you are intentionally powdering the sample then I would assume you want reliably measure the bulk properties of a sample much thicker than 200 nm. In which case transmission would indeed be the best option as it will pass through the sample, providing information on its full depth and exposed cross section.
ATR is good for measuring surface properties, but generally not great for bulk ones. It will provide accurate estimates of spectra from the local exposed area (as Muzammil mentions this is not very deep), but not of the bulk, meaning that its sampling error is large wrt bulk properties.
I have prepared different powder samples from one raw material using different parameters such as temperature and time. Then. I want to check their difference on functional groups, intensities and others.
In contrast to ATR, the Beer-Lambert law works quite well if you use the KBr (or CsI) technique and individual grains or crystallites are small compared to the wavelength (i.e. smaller than about one tenth of the wavelength). See e.g. Article Employing Theories Far beyond Their Limits-The Case of the (...
Water content might be some problem, but now you can do something against it: https://www.ncbi.nlm.nih.gov/pubmed/28738626
Attenuated total reflectance (ATR) is a special accessory unit which can be used with Fourier transform infrared (FTIR) spectrometers. It enables you to measure directly onto a solid state sample surface by pressing the sample towards an ATR crystal (e.g. diamond), including liquid and gas samples, thus avoiding the need to prepare pellets (small concentration of sample in e.g. KBr, mechanically and/or chemically dissolved) for use in the normal transmittance mode. The bulk of a sample is readily measured by first slicing off the top layer of the sample to the desired depth (as long as that is possible). The ATR-FTIR technique makes it possible to study materials which are non-transparent to infrared radiation in a pristine condition. Hence, the extensive, time-consuming and often cumbersome sample preparation by pressing thin KBr pellets as in traditional FTIR transmittance spectroscopy is avoided. The traditional technique might even change the sample material in question. The ATR technique is based on a special reflectance setup where the sample is pressed directly onto various crystals with high refractive indices, e.g. diamond. When interpreting ATR-FTIR spectra where quantitative results are important, note that non-corrected ATR spectra have much stronger absorbance bands at longer wavelengths (smaller wave numbers) than at shorter wavelengths (larger wave numbers) compared to normal FTIR transmittance spectra. The reason being that the penetration depth is dependent on (in addition to refractive indices of sample and ATR crystal, and angle of incident radiation) the radiation wavelength, and increases with increasing wavelength (decreasing wave number). The sample should be measured at different locations (where you can specify the FTIR instrument to run several spectra in one run, e.g. 32 scans, which are automatically averaged into one spectrum), but since you’re doing ATR-FTIR you should not make an average of the spectra taken at different locations with different contact variations between the sample and the ATR crystal. Relatively hard surfaces complicate accurate quantitative measurements (height of absorbance peaks) due to varying contact with the ATR crystal for the different samples. Air between sample and ATR crystal results in a weaker absorbance signal. Thus, unless other conditions indicate otherwise (e.g. inhomogenities, impurities, etc.), the FTIR spectrum with the largest absorbance peaks represents the most correct measurements on one and the same sample, and hence this spectrum should be chosen as it is assumed to be the most correct one. Qualitative measurements (location of absorbance peaks at wave numbers) do not represent a problem as long as the contact area is large enough to ensure a sufficient strong measurement signal. Various details about the ATR-FTIR technique which may be helpful may be found in the following article (may be requested through Research Gate): B. P. Jelle, T.-N. Nilsen, P. J. Hovde and A. Gustavsen, ”Accelerated Climate Aging of Building Materials and their Characterization by Fourier Transform Infrared Radiation Analysis”, Journal of Building Physics, 36, 99-112, 2012.