Reversed flow where it shouldn't
 

Hey guys,

I'm stuck with my simulation and need some help.

I want to simulate flow through a manifold, to get a qualitative result of what the flow might look like.

The manifold is part of a measurement system and i can briefly explain what the flow should look like in there.

We have the main-flow from inlet 1 to outlet 2 due to a delta p which I dont know exactly, I just now that this flow is around 1,5 l/min.
Outlet 5 is the flow to the detector, which is held constant at 120ml/min by a pump and a mfc outside of the fluid domain.
Now the middle one of the three inlets shown in the picture is a special one and actually the one I'm interested in. This inlet is 0,75mm ID pipe and it has a 0,375mm OD/0,03mm ID capillary in it. The outer pipe is an outlet (pump and critical orifice outside flow domain) with around 10ml/min flow and the capillary is an inlet with 0,2ml/min flow. This results in the fact, that the fluid going in through the capillary gets immediately sucked away by the outlet 7, which is intendet.
The fluid through inlet 6 is NO2, whereas the fluid through all the other inlets air is.

http://www2.pic-upload.de/thumb/31892335/manifold-4.jpg



After this simulation I want to see how deep does the NO2 get pushed in the manifold (y-direction) before it gets sucked away by outlet 7. I assumed I can use the species transport model with an inert-mixture (because there is no chemical reaction between NO2 and air) to get to this result.

Now my problem is: when I set outlet 7 as pressure-outlet with target mass flow I get reversed flow in this surface. How do I get rid of this?

Some information on my setting: