Are you trying to crystallise something by dissolving it in a good solvent and then layering that with a poor solvent so that they slowly diffuse into each other and force your compound out of solution?

First, make sure that your two solvents will mix with each other. Layering water and chloroform won't work.

Choose solvents that have different densities where possible, so that it is easier to keep the layers separate while you set up your experiment. Methanol and ethanol will mix very quickly, but methanol and chloroform will mix more slowly.

Put your less dense solution into the vial first. Then, use a disposable glass pipette as a funnel to pour your dense solvent to the bottom of the vial. Make sure that your funnel pipette is touching your vial base. You can slow the flow of liquid by placing your thumb or finger on to the top of the funnel pipette.

I have only used this method to make compounds, so I would have a metal salt dissolved in one layer and my ligand dissolved in the other.in these cases, it helps to have a buffer layer of one of your solvents (without anything dissolved in it) between your two reagent layers. If you only have something dissolved in one layer, I would use a buffer layer of that solvent. This stops the system from having a layer of solid form at the interface of your two layers.

If you have trouble, let me know

Laura_Mccormick   first of all thank you mam for your kind information. Actually I al ready prepared the metal ligand complex. Now I am facing a problem on recrystallizing of the product.Some information I got that it can be recrystallized by applying double layer solvent diffusion method. So my question is about choosing those two solvents. Could you kindly help me about that problem ? or is it possible to tell me some convenient method to re crystallize the product.

The first step would be to screen solvents. Dropping a small amount (~1 mg)  into ~1 mL solvent and looking for mixing lines followed by waiting to see if it dissolves with time and perhaps sonicating to break up particles. The solvent pairs that have worked well for me are acetonitrile or acetone (soluble) under ether (insoluble). With aromatic ligands, consider toluene/acetonitrile or p-xylene/acetonitrile. If your complex is soluble in chloroform but not methanol, that is another possible option.

For initial screening, you can use 5 mM glass tubes made from glass tubing (not expensive NMR tubes). You only need about 1-1.5" of the complex solution - the rest is filled with the cosolvent. This can be a good way to screen solvents since it uses a rather small volume. Remember that adding a compound to a solvent is likely to increase its density and you need a reasonable difference for layering.

For larger scale recrystallizations I use 4 mL vials with 1-2 mL of the more dense solution (10-20 mM is typically a good starting point)  in the bottom and use a blunt tip syringe to dispense the less dense solvent SLOWLY dropwise into the center of of the meniscus. The syringe helps mitigate the tendency of volatile solvents to creep/drip (as does the method recommended first). Once the top layer is well established, ether can be added down the walls of the vial. Stop/go MUCH slower as soon as you see sustained precipitate. Vials work well when the insoluble solvent is less dense, I normally stick with narrow tubes when the less soluble solvent is more dense (eg. toluene/acetonitrile). Hint: mark the interface and the total volume with a Sharpie to help follow what is happening as time passes - especially with tough to retain ether. 

well I am facing a similar problem, I have used 4 ml vials, dissolved my metal complexes in dichloromethane, and layered them with hexane. they did not mix in the beginning, and after leaving them for few days, they have all precipitated powdered solid. As far as I know that requires decreasin the concentration of the initial solution (I started with 1.0mM solution)

Now my question is, should I decrease the concentration only, or also try different solvent mixes (acetonitrile/diethyl ether)