Since Liu and Cohen predicted theoretically an unknown metastable carbon-nitride compound β-C3N4 with bulk module and hardness similar to those of diamond [1, 2], many groups have tried to synthesize this kind of carbon-nitride material on various substrates such as Si, Ge, NaCl, KCl, Ni by a variety of film fabrication techniques [2–6]. Most authors reported the formation of amorphous CNx with x < 1. Nevertheless, some authors have reported that crystalline β-C3N4 or α-C3N4 with sizes of five to hundreds of nanometers have been observed in amorphous CNx films [2, 3]. Wang et al. [7] and Tani et al. [8] have reported that they have successfully synthesized β-C3N4 with approximately stoichiometric composition with an atomic ratio of N/C= 1.3–1.4 by using chemical vapor deposition and electron cyclotron resonance plasma sputtering methods, respectively. Meanwhile, some other C-N compounds with structures that have not been predicted theoretically were also observed [5, 9–11]. These results indicate that up to now the exact structures of the C-N compounds are still not clarified. Recently, Teter and Hemley [12] have reinvestigated the stability and properties of carbon-nitride using first-principle calculations. They found that carbon-nitride compounds have five possible metastable structures: α, β, cubic, pseudocubic and graphite-C3N4. In these structures, the metastable cubic C3N4 (which might have the highest zero-pressure bulk module of 496 GPa exceeding that of diamond 458 GPa), and the graphite-C3N4 are energetically preferred over the other four structures. Another important piece of information from their calculation is that the graphitic C3N4 is a possible precursor for the synthesis of βor cubic C3N4. The transition pressure from graphite-C3N4 to cubic-C3N4 is only 12 GPa according to their calculation. This implies that cubicC3N4 can be synthesized from precursor graphite-C3N4 on an industrial scale similar to the synthesis of diamond or cubic-BN. Therefore, it will be particularly important to find proper methods for synthesizing larger graphite-C3N4 crystals. Among the predicted structure types of graphite C3N4, two are of space groups P 6m2 [12] and R3m [13] which are stacked according to the sequence of ABAB · · · and ABCABC · · ·, respectively. In the Teter and Hemley predictions [12], the vacancies ordered in the graphite plane can also lead to two new graphite phases of space group P2mm with the stacking sequences of AA · · · and ABAB · · · [14]. Table I lists the structural parameters of four possible graphite C3N4. The synthesis of a metastable phase depends not only on the thermodynamic condition but also on the dynamic process. Choosing an appropriate material as the structural template for the nucleation of a metastable phase to overcome its energy barrier has become a valid method in practice. Li et al. [15] made a first test to produce crystalline carbon nitride using CNx/TiNx multilayers. TiN was chosen as a template material because of the good lattice match between TiN (111) and β-C3N4 (001). TiN (111) planes have hexagonal symmetry with a unit vector of 3.0 A, the lattice mismatch between these two planes is about 7%. It should be noted that two unit vectors (=9.48 A) of graphite-C3N4 (0001) would fit better into three unit vectors (9.0 A) of TiN (111) with a lattice mismatch of 5%. Therefore, we reasonably believe that graphite C3N4 can form more easily in the CNx/TiNx multilayers than β-C3N4. We have successfully prepared the C-N films with β-C3N4 with a size of a few nanometers by using an ion beam sputtering method, details are shown elsewhere [6]. In this paper, we present results on the structural characterization of the graphite-C3N4 in the CNx/TiNx multilayers by selected area electron diffraction (SAED) and X-ray diffraction. The TiNx/CNx multilayers were prepared by using an ion beam sputtering apparatus. The targets were graphite and titanium and the discharge gas was