I wonder, if you could enclose your spectrum, the comments might be much more focused.

Do you have a COOH group in liposomes? The corresponding carboxyl band at ~1740 1/cm is very sensitive to environment, and can be shifted from 1715 (very strong H-bonding) to 1760 (very little external influence).

There was a nice paper by Aihua H. Xie:

BIOPHYSICAL JOURNAL, Volume: 88, Issue: 4, Pages: 2833-2847 (2005)

DOI: 10.1529/biophysj.104.047639

Abstract

Hydrogen bonding is a fundamental element in protein structure and function. Breaking a single hydrogen bond may impair the stability of a protein. We report an infrared vibrational spectral marker for probing the hydrogen-bond number for buried, protonated Asp or Glu residues in proteins. Ab initio computational studies were performed on hydrogen-bonding interactions of a COOH group with a variety of side-chain model compounds of polar and charged amino acids in vacuum using density function theory. For hydrogen-bonding interactions with polar side-chain groups, our results show a strong correlation between the C=O stretching frequency and the hydrogen bond number of a COOH group: similar to 1759-1776 cm(-1) for zero,; 1733 1749 cm(-1) for one, and 1703-1710 cm(-1) for two hydrogen bonds. Experimental evidence for this correlation will be discussed. In addition, we show an approximate linear correlation between the C=O stretching frequency and the hydrogen-bond strength. We propose that a two-dimensional infrared spectroscopy, C=O stretching versus O-H stretching, may be employed to identify the specific type of hydrogen-bonding interaction. This vibrational spectral marker for hydrogen-bonding interaction is expected to enhance the power of time-resolved Fourier transform infrared spectroscopy for structural characterization of functionally important intermediates of proteins.

The article suggested by Prof Andrei Dioumaev gives clear answer to your question.  

As you are comparing the C=O stretching of a component in your study system, in presence of various polymers and liposomes,  the reference is the C=O stretching wave number (reference wavenumber) of the single component, where there could be  inter / intra molecular hydrogen bonding existing between the molecules of same species. Upon exposing the single component to other components, the added components can either introduce more intra molecular hydrogen bonding (lower wave number) due to hetero molecular interaction on the C=O group, or the added components can inhibit the inter / intra molecular hydrogen bonding which existed in single component (higher wave number). 

If bond length decreases the peak wave number shift to higher values.  If it decreases the bond length increases.  Bond length changes may occur due to the change in electronegativity of the neighboring atom.  This is something like hydrogen bonding.

Can you upload your IR spectrum

Everything said, just be careful: C=O is a good tracer for the protonated/deuturated  form of carboxylic acid, COOH & COOD. It is NOT the CO-H stretch frequency  but rather the antisymetric stretch of COOH, essentially the C=O stretch that is strongly coupled to the C-OH and/or C-OD stretches (with a 8-12 cm-1 downshift, i.e. a"red" shift, in COOD vs. COOH). If you put a bulkier group COO-R, methyl  or ethyl, instead of the proton (e.g. -COO-CH3), the coupling is practically lost, and the C=O stretch frequency is losing (some of) its sensitivity to the changes around the adjusting C-OR stretch.

To answer the question properly, it is of vital interest to know what measurement technique was applied, since band shifts can also be caused by optical phenomena without any change of the chemical structure. E.g. polymers can be investigated in the form of freestanding films. In such films interference effects can lead to band shifts (and even to the occurrence of bands which cannot be attributed to vibrations), see e.g.Article The electric field standing wave effect in infrared transfle...

Article The Electric Field Standing Wave Effect in Infrared Transmis...

Also, preferred orientation can lead to band shifts as well as crystallite size effects, see e.g. Article Employing Theories Far beyond Their Limits - Linear Dichroism Theory

Such effects must be properly corrected before any "chemical" interpretation of band shifts is possible.

Please refer bathochromic shift

Bathochromic shift: In spectroscopy, the position shift of a peak or signal to longer wavelength (lower energy). Also called a red shift.

If FTIR peak shift is only 1-5 cm-1, can we define this as a real shift or not? thanks for whoever answer this question.

Fangjie Qi What do you mean be "real"? One might define it relative to the spectral resolution you use - on the other hands, with 2D correlation spectroscopy you can detect shifts much smaller than the spectral resolution. If you mean by "real" chemical vs. optical reasons, this depends on your measurement technique. ATR peaks, e.g., are generally redshifted. I recommend in this regard the following review: Article The Bouguer-Beer-Lambert Law: Shining Light on the Obscure

I think this can be correlated to the bond length

There are a multitude of different explanations for band shifts apart from changes in the bond length, which we cannot understand from the viewpoint of the Beer-Lambert approximation. You have to go one step further to understand them properly:

Preprint Wave optics in Infrared Spectroscopy