Dear Truong-Giang Vo,

Higher PL intensity indicates less nonradiave defects' density(can be surface states)  when the compared structures are the same ( for ex: same material with different growth conditions). 

the increase of the PL peaks intensity is due to an increased radiative recombination probability over nonradiative one (collection of more generated photons)

Bests 

One possibility:

Normally surface states are due to defect. Large surface area may indicate more surface defects which will quench the emission from the bulk ZnO.

AS per my understanding, if you are talking about the NBE (CB to VB transition) then your first statement  i.e higher PL intensity for the sample having less surface states (defects) and better crystallinity is correct.

Dear Mallick and Dr. Zhao,

Thank you so much for your kind help. 

If higher PL intensity implies a lower density of surface state, is it contradictory with the hypothesis claiming that higher PL intensity. the higher recombination between electron and hole.

Best regards,

Dear Truong-Giang Vo,

Higher PL intensity indicates less nonradiave defects' density(can be surface states)  when the compared structures are the same ( for ex: same material with different growth conditions). 

the increase of the PL peaks intensity is due to an increased radiative recombination probability over nonradiative one (collection of more generated photons)

Bests 

Dear Truong-Giang Vo,

I could give you an explanation on example of quantum dots.

If you have your PL inside the core of a quantum dot and quantum confinement governs the PL process, then normally high crystallinity is required and the absence of surface defects (dangling bonds) thanks to good surface passivation.

However, sometimes electron or hole or both are not quantum confined or partially quantum confined only. Then you could expect that surface luminescence is dominant. If you eliminate surface defects - you quench your PL.

Hope it was helpful.

Cheers!

The PL intensity we are talking about is band to band excitonic recombination i.e. an electron at the CB to a hole at the  VB recombine.

So it is not contradictory to the hypothesis toy are talking about.

For the case of larger defects at surface /within the crystal, the band to band transition reduces at a cost of enhanced visible emission means the electron from the conduction band either radiatively recombine with the hole trapped in a defect level or vice versa.

Please see the attachment. I think it  will be clear to you now.

Here doped sample has more defects that is why it has reduced the NBE intensity.

Dear all,

I greatly appreciate for your enthusiastic helps that make my confusing issue clear now. 

Wish you all the best.

Sincerely yours,

Aridam Mallick is right. One has to specify which emission band is of interest. Also, it is quite common that a lower crystallinity sample shows over all stronger PL, but not from the bandgap, because a localized state, if not nonradiative, is less affected by the nonradiative defects. InGaN LED is a good example.  

First we need to identified the PL, coming from band edge or surface state, if band edge then defaces may decrease the intensity.

The intensity of the peak decreased as the film thickness increased being strongly dependent on the grain surface area. In order to well define band gap or excitonic emission, it is necessary to control the surface states. In general, the density of the surface states decreases with the grain size increase. Also, luminescence intensity decreases with decrease of the density of the surface states due to small surface to volume ratio for the larger crystallites. Thus, the films with smaller crystallites could show higher luminescence intensity compared to the larger crystallites.

read more on this paper:

Nihan Akin, Yunus Ozen, H. Ibrahim Efkere, Mehmet Cakmaka, and Suleyman Ozcelika,

(wileyonlinelibrary.com) DOI 10.1002/sia.5677