Mr. Chatterjee,

First of all what is the resolution of your FTIR instrument? You have shown in your figure the frequencies up to second decimal sign? It those values are within the frame of the resolution of the spectrometer, because of if they are not than there is the second question are both values 1530.14 cm-1 and 1531.34 cm-1 distinguishable?

If you have operated with a become standard resolution for many FTIR instruments of +/- 0.5 cm-1, for example, than it is no possible unambiguously to conclude that you have meaningful shifting of your IR-bands as they are shown.

Secondly, if you have performed automatic peak picking, usually the shown values do not corresponds to the real, ones due to a broad and especially in your case, overlapped character of IR-patterns.

In such cases it is important to perform chemometric analysis, involving deconvolution and curve fitting procedures in order to determine: (i) number of the sub-maxima, (ii) their position and (iii) intensities.

Please pay attention to few of our papers in topic "quantitative vibrational spectroscopy" [Refs. 1-3].

Thirdly, to evaluate even qualitatively a change of IR-patterns due to any kind of process such as for example, adsorption, chemical equilibrium between different forms in solution and more, especially like your systems where IR-profile is broad, overlapped and there are not expected significant shifting of IR-modes, it is important to use also spectroscopic subtraction procedure. It consists on, in your case should be, a subtraction of IR-spectrum of Au-nano-particles from this one of corresponding adsorbed system. The resulting spectrum should correspond to IR-spectrum of adsorbed dipeptide system, which presumably should be perturbed due to interactions onto Au-surface. Than this spectrum you can compare with an IR-one of non-perturbed system. Thus you can easy evaluate (even to quantify), which frequencies are affected due to coordination to Au-surface, and how much they are perturbed. Towards subtraction procedure, please pay attention to [Ref. 4]. Those chemometric approaches you can apply, using large number of statistical software available.

[Ref. 1] Koleva, B.; Kolev, T.M.; Spiteller, M. (2008) Determination of cephalosporins in solid binary mixtures by polarized IR- and Raman spectroscopy, J. Pharmaceut. Biomed. Anal. 48, 201-204.

[Ref. 2] Lamshoeft, M.; Ivanova, B.; Spiteller, M. (2011) Chemical identification and determination of sulfonamides in n-component solid mixtures within THz-region-solid-state Raman spectroscopic and mass spectrometric study, Talanta, 85, 2562-2575.

[Ref. 3] Ivanova, B., Spiteller, M. (2012) Quantitative solid-state Raman spectroscopic method for control of fungicides, Analyst, 137, 3355-3364

[Ref.4]. Ivanova, B.; Tsalev, D.; Arnaudov, M. (2006) Validation of reducing-difference procedure for the interpretation of non-polarized infrared spectra of n-component solid mixtures, Talanta 69, 822-828.

I agree with Prof. Ivanova.  Broad bands  (possibly coming from water on the BSA) could easily shift your apparent peak position by 1 cm^-1.  It looks like after conjugation with gold, the BSA spectrum looks free of water peaks so that you have better isolated peaks that are easier to fit to a curve.  The spectra with bands coming from possibly water is harder to fit to curve since such peaks  (in liquid states) tend to be broad and asymmetric. 

i have the

same thing but one of my band of BSA at 1307 disappear, what does it mean please?