This tutorial runs a chamber cleaning recipe.

Step 1: Safety Check

Read the safety instructions on the equipment.

Check the alarm system. There should be only the power light among all lights on the top. Otherwise there may be something wrong.

Also notice that there is a chloride gas detector on the side wall. If that detector alarms, immediately get out of the clean room and notify staffs.

Step 2: Initializing

Open the Oxford program on the desktop, and log in. Click SystemPump to switch to the Pump view, which shows a basic system overview: loadlock, chamber, vacuum pump and nitrogen fixture. There are also several sub-windows showing loadlock gas pressure, chamber gas pressure, the movement of trail chain between loadlock and chamber, etc.

Figure 1: Pump view

pump view

Usually the chamber gas pressure should be below 5e-5 Torr before venting.

Now open the nitrogen valve.

Always check if the nitrogen valve is open before running the recipe.

Now it is recommended to set the temperature at first. Go to ProcessRecipe, select a recipe (any recipe works in this case), and set the temperature (left bottom of Figure 1). Click the start, and then stop shortly after that. This will bring the temperature to the point you want.

This is how to set chamber temperature. This trick is also employed when signing off in Step 7.

The reason to do this is mainly because of the 10 degree C temperature tolerance of the system, and that temperature variation is unavoidable. For example, consider the case in which the target temperature is 10 degree C and the current temperature is 50 degree C. If one runs a recipe without presetting the temperature, the process may wait until the chamber temperature goes down to 20 degree C (instead of 10 degree C because of the 10 degree C tolerance). But in real world the temperature fluctuates. It is possible it goes back to 21 degree C after first hitting 20 degree C, and then back to 19oC and so on. The process may be terminated and restarted back and forth due to this fluctuation around 20 degree C. This is not good for the equipment, and may have a negative impact on the sample as well.

Step 3: Venting and Loading

In the loadlock subwindow, click VentStopVent. Wait for about 2min until it finishes.

The system has a glitch: if you just log on and vent the chamber for the first time, you should do VentStopVent. After that, only StopVent is needed.

Open the loadlock and load the carrier. Make sure the round edge of the carrier touches the two pins.

In the loadlock subwindow, click StopEvacuate to pump down. A diaglog box will pop up asking for the wafer name. Since we have a carrier inside, type in a name you want (such as “Si carrier”). You can see a green spot appears on the loadlock. It means it is busy (pumping), which may take about half a minute.

After the evacuation is done, there will be green arrows in the left upper corner of the chamber and loadlock, meaning they are ready for further steps.

Step 4: Selecting Recipe

While the loadlock is being vented, click ProcessRecipeLoad and select the recipe. In this tutorial we select the recipe called “Chamber O2 RIE only clean 40dcg” to clean the chamber.

There are multiple steps in one recipe. To view or edit the each step, left click on the step and then Edit.

Step 5: Running Recipe

Go back to the recipe window (ProcessRecipe), and press Run to start. If you go back to the pump view (SystemPump), the green arrows in the left corners of loadlock and chamber will be gone, meaning these two places are busy. You can see the valve is open (carrier being transferred to the chamber from the loadlock), and then a square appears on the chamber (meaning the carrier is to the position).

Back to the recipe view, there are several parameters that need special attentions.

In the RF Generator section, make sure the Reflected Power is below 10% of the Forward Power.

In the Helium Backing section, the flow meter should be below 10 Sccm. The flow meter measures the leaking helium gas pressure. The helium here is used to cool the carrier down from the bottom. If it is too large (e.g., above 10 Sccm) there must be some problems (e.g, carrier off the center, valve not fully closed, etc).

Step 6: Logging parameters

Open the lab notebook besides the chamber, and write down some important parameters, such as O2 pressure, RF forward power, reflected RF power, pressure, temperature, helium pressure, leaking helium pressure (flow meter), time, DC bias, ICP power, etc.

In the middle of the recipe, one can Pause, Stop or Jump. Pause shuts off the active power (RF power, ICP power, and such), but keeps the gas flowing. It just temporarily suspends the step in a recipe. Stop terminates the whole recipe. If you find yourself loading a wrong recipe, do that. Jump skips the current step and continue with the next step.

Stop versus Jump is just like break vs continue in C programming language.

Step 7: Finishing

In recipe view (ProcessRecipe), make sure nothing is flowing (gas flow zero, RF power zero, etc). Set the temperature to 50 degree C, and click Start, and then Stop shortly after that.

We don’t really want to run a recipe, but just reset the temperature. By pressing Start, the system tries to bring up a recipe run, including setting the temperature. We press Stop shortly after that to abort this action because we don’t want to actually, but the temperature will still be dragged back to 50 degree C regardlessly.

Vent the chamber (in chamber view, StopVent), unload the wafer and carrier, evacuate the chamber, and exit the program (SystemExit).


  • Sometimes after you log in you may see the system is not running (e.g., in the Pump view, there is no "smoke" coming out of the pump, the engine is not rotating, etc.). It is not necessarily the case, since it may be the communication failure between the software and the equipment. Try to log out and log in again (reinitializing the communication). If this does not work, try to tweak the BNC connector on the back side of the computer for a little bit. If this still does not work, the system may be really out of order somehow (rare…).

  • Sometimes the reflected RF power is more than 10% of the forward power. This may still be okay. Check the "REV" value on the little metal box in lower level. If this shows fine, then it should be fine.