It isn't. The work function of any material is not temperature dependent.

This is subject of study, I think, especially for poly-C materials. For metals the dependence is linear:

http://link.springer.com/article/10.1023%2FA%3A1017556031226

http://journals.aps.org/pr/abstract/10.1103/PhysRev.59.75 

For semiconductor WF variation not so evident and depends on doping level.

You may have alook: Semiconductor Surfaces and Interfaces by Winfried Mönch

for details.

Dear Ravi Kumar,

Dr. Lawrence is right that work function isn't temperature dependent. However, band-to-band transitions are affected by surface states on silicon (doped with any dopant) which create a difference . There is a shift of band-to-band absorption from 1.04ev (crystalline silicon) to 1.18ev (surface of silicon) which in turn result in band tilting at surface. Interestingly this band tilting at surface is drastically affected by low temperature annealing.  If we compare poly-Si with crystalline Si, then poly has more surface area than crystalline; similarly amorphous has more surface area than poly-Si. The more surface area, the more surface states.

Another interesting result is that in most cases silicon surfaces are contaminated by Carbon that doesn't stack at substitutional sites, rather form a non-periodic lattice structure of SiC. The structure thus result in band-tilting and in turn affect the work function at surface.

For details please see my two papers below;

1) Japanese Journal of Applied Physics, Vol. 43, No. SA, 2004, pp. 2397-2401

Effect of Surface States on Piezoelectric Photothermal Specta of Silicon Single Crystals

Aftab A. MEMON, Miroslaw MALINSKI1, Atsuhiko FUKUYAMA and Tetsuo IKARi

2) “Chemical contamination of thin oxides and native silicon for use in modern device processing”, V. SUBRAMANIAN, P. K. BHATTACHARYA & A. A. MEMON, International Journal of Electronics, Volume 78, Issue 3, pp: 519-525 (1995), ISSN: 0020-7217