If your sample contains Al element, it is not suitable to deposit it on the Al foil and use Al as a reference for calibration. Because Al 2p peak in Al metal locates at 72.7 eV and that in Al2O3 locates at 74.7 eV (if similar to your sample), difficult to be distinguished. Please use copper. 

I am more concerned about the oxidation state of Al. Al can be spontaneously oxidized upon exposure to the atmosphere. In this case, your reference may not longer be accurate (It could be partially oxidized as well, how would you know these details). Even using Cu foil, can you be sure the chemical states Cu(0) or Cu(I) or Cu(II)?The binding energies of metallic Cu and oxidized Cu are very close to each other.

Most people use carbon or other noble metal (Au & Ag) to do characterization as these values are well known and easily being identified (spin-orbiting splitting, i.e., the peak separation and area ratio).  Even using internal Ag as the reference, Ar sputtering is often carried out to remove the oxide layer (though very insignificant). I have experience operating XPS, this is a standard procedure for calibrating a XPS machine monthly.  

I strongly suggest you to use adventitious carbon as the reference. Assuming you know the Al chemical state well, it is always no harm to cross check with adventitious carbon. 

C 1s peak can be deconvoluted to (a) aliphatic C-C & C-H, (b) C-OH & C-O-C, (c) C=O and (d) O=C-O. This means that adventitious carbon may contain oxygen in nature. 

After deconvolution, set aliphatic carbon C1s to be 284.8eV.

For polymers, some people prefers to set it as 285.0eV. Hope it helps. 

 If you are using air exposed powder you should use Indium foil and press the powder slightly in a previously with a scalpell scratch cleaned part of the foil. In d levels, 3d and 4d, serve as a good  refernce. 

Binding energy referencing in XPS is a rather complex issue. First, you  need to make sure that your specimen makes a good electrical contact to the spectrometer. If, for example, you deposit on oxidized Al this condition won't be satisfied and energy levels of your sample can be off.

There are also issues with using adventitious carbon, see good review by T.L. Barr, S. Seal, J. Vac. Sci. Technol. A 13 (1995) 1239

Also if there's a flood gun on the system you might use it rather than doing a charge compensation if you want to be sure the readings are quite accurate, but Al foil should make a nice contact with the sample holder so it shouldn't be a (significant) charge shift in your measurements.