Phosphorous diffusion is used to introduce an n-type layer on the surface of a p-type wafer. The formed p-n junction acts to collect light-generated carriers so a current can flow in an external circuit. The n-type layer is commonly referred to as an emitter and the concentration of electrons in that layer depends on how the phosphorus diffusion is performed. PV Factory simulates an inline diffusion process. The properties of the emitter are determined by the concentration of the phosphorus source, how fast the belt moves through the furnace and the temperature in the furnace. Before the tutorial, read the help files associated with the phosphorous diffusion process to familiarise yourselves with the process.
- Explain why you form a p-n junction to collect electrical carriers from a silicon solar cell
- Explain the benefits and disadvantages of inline diffusion as compared to tube diffusion
- Be able to perform a main factor experiment to determine the most important parameter to optimise
- Be able to perform a single factor experiment to optimise the acidic texturing process
- Understand how to characterise an emitter
Please review the PVmanufacturing pages phosphorus diffusion and the help files in PV Factory before attempting the tutorial so that you understand the effect of the different parameters. For this tutorial exercise you will use the Factory Default settings for all processes following the Phosphorous Diffusion step so leave your UNSW Screen-Print line settings to “Use factory default”. All experiments should use batches of wafers (at least 10 per batch) with the following properties:
- Standard Cz mono-crystalline silicon wafers;
- 200 mm thick;
- Resistivity of 1 W cm; and
- Cut using a standard wire saw (this assumes a wire saw that uses a slurry not diamond tips).
In this exercise, you should use your best recipes that you have established for Saw Damage Removal and Alkaline Texturing. To load these recipes use the “L” icon on the left-hand screen when you are processing that step for batches. Otherwise, you can set your user settings to “Use my recipe”, provided the steps following phosphorus diffusion are PVfactory default. When you get to the In-line Phosphorous Diffusion step, enter your settings, and run that step. Immediately after the diffusion has been performed, you can check the average sheet resistance of the diffused wafers. To determine the effect of your diffusion process on your cell efficiency, select Run Remaining Steps. This will cause your batch to be processed using the Factory Default parameters for all the remaining steps (because you have not established Best Recipes for these steps yet). The I-V results for your batch will then be available.
You can then go into the Measurement Lab and do an Electrochemical Capacitance-Voltage (ECV) measurement to obtain a dopant profile for a selected wafer, but be warned you will break a cell to do this measurement as it is a destructive process. Consequently, it is best to complete processing your batch before you perform this test. In order to determine the junction depth for your cell, you will need to work out what p-type dopant concentration corresponds to a wafer resistivity of 1 W cm. Best to work this out using this PVLighthouse calculator .
So the procedure you will need to follow this week is:
Create New Batch (mono-crystalline wafers) → Saw Damage Removal (best recipe) → Alkaline Texturing (best recipe) → HF/HCl clean → Phosphorus Diffusion (your experiment) → Return to Phosphorous Diffusion step and record your average sheet resistance → Complete Batch and download I-V data → Go to Characterisation Lab (ECV) and record junction depth.
Part 1 – Main Factor-Response Experiment
As mentioned above the diffusion process depends on many parameters including the phosphorus source concentration, belt speed and in-line furnace temperature. In general, in-line furnaces have a number of heating zones with the earlier and later zones being used to ramp up and ramp down the temperature so that the wafers are not stressed by rapid changes in temperature. In PV Factory, the peak temperature is only set for Zone 2 and therefore this zone is the most critical heating zone where most if not practically all the diffusion occurs. As with acidic texturing, variations in belt speed reflect different diffusion durations as the length of the furnace (and belt) is fixed and depends on the in-line tool.
Design a main factor response experiment which explores the effects of the following factors of interest (allowed ranges in PV Factory are in parentheses):
- Phosphorous source concentration (10 – 100 % (v/v));
- Belt speed (0.1 – 1.0 cm s-1)
- Zone 2 temperature (700 – 1000 ºC)
on the responses of surface concentration (cm-3), junction depth (μm), sheet resistance (Ω/square) and mean cell efficiency (%).The suggested factory settings are listed in the table below, but you can also select other settings to explore which factor is most significant.
|Factor Settings||Main Factors for In-line diffusion|
|Zone 2 Temperature
Main Factor Response Experiment: Activities and Questions
- If you are completing this tutorial as a class, get your tutor to check your experimental design.
- What is your base dopant concentration, NA?
- Before you start this experiment decide whether you need to process 10 batches to completion. Hint: How many batches have you processed in your saw damage and alkaline texturing baths? Why?
- Then process batches for each of your designed experiments and record the results.
- Produce a main factor response graph for each of the four responses.
- Which of the factor is most significant in determining cell efficiency?
- Of the less significant factors can you observe trends which suggest an optimum value?
Part 2 – Single Factor Response Curve
Identify the factor that you think has the greatest effect on the batch mean cell efficiency and run a single factor experiment to determine your preferred setting for that factor. We’ll also record the other responses that you considered in Part 1.
Single Factor Response Curve Experiment: Activities and Questions
- Use the same wafer parameters as you used for Part 1.
- Create and process at least 8 experimental batches (with at least 10 wafers per batch), varying the factor of interest over the range of values allowed by PV Factory. If you note some interesting trends then perhaps consider sampling additional factor settings so that you can clearly observe any trends.
- Record all responses for each factor setting.
- Sketch an X-Y scatter plot for each response (y-axis) versus your factor of interest (x-axis).
- Describe the relationships between your factor and each of the responses.
Part 3 – Understanding In-line Phosphorus Diffusion
The following questions are used to determine your understanding of the phosphorus diffusion process. You may prepare your answers while completing the above tasks.
- Describe two ways in which the phosphorus dopant source can be applied to the wafer before in-line phosphorous diffusion.
- What are the benefits of using in-line diffusion compared to performing a diffusion process in a POCl3 furnace?
- How do dopant profiles measured using SIMS and ECV differ?
- Figure 1 shows ECV profiles of 4 different emitter, all having a Rsh of 70 Ω/square. Label the emitters on the graph: (A) Etched back from 40 Ω/square; (B) Etched back from 50 Ω/square; (C) Etched back from Ω/square; and (D) Diffused as Ω/square.
- If you do an etch-back after diffusion on mono-crystalline wafers, do you remove all the pyramids?
- What is a major limitation of the optimisation that you performed today?
- A heavily-doped emitter is needed in screen-printed solar cells to improve the blue wavelength response.
- A 200 μm-thick layer with a resistivity of 1 Ω cm gives a sheet resistivity of 100 Ω/square
- A lightly-doped emitter introduces low resistive losses.
- An infinite doping source ensures that the phosphorus concentration at the surface remains constant.
- Gettering which occurs during phosphorus diffusion helps reduce recombination in the wafer.
- With inline diffusion, phosphorus only diffuses into the front surface and so you do not need to edge isolate the emitter.