Abstract: The advantages of low-pressure phosphorus diffusion over atmospheric pressure diffusion are analyzed, and the working principle and adjustable process parameters of low-pressure phosphorus diffusion are discussed. By studying the influence of diffusion process parameters such as source flow rate, diffusion time, source bottle pressure, vacuum pump pressure, etc. on the diffusion block resistance under vacuum conditions, understand the debugging ideas, ways and methods of the low-pressure phosphorus diffusion process, so as to facilitate the development of low-pressure phosphorus diffusion process more efficiently debugging.
introduction
With the rapid development of the global economy and society, coal, oil and natural gas-based fossil energy sources are increasingly in short supply, environmental pollution and ecological deterioration have deepened, and new clean energy has attracted much attention. Solar energy has become one of the most potential renewable energy sources due to its rich resources, wide distribution, easy utilization, and low cost. Solar power generation technology has matured to meet the growing demand for electricity worldwide. The solar cell is the most important part of the photovoltaic power generation system. The p-n junction is the core part of the crystalline silicon solar cell. Without the p-n junction, you cannot convert light into electricity, and you cannot become a solar cell. Therefore, the manufacture of p-n junctions is the most important process.
Tubular diffusion furnace, especially horizontal diffusion furnace, is one of the important process equipments in the diffusion process of photovoltaic and semiconductor production lines.
At present, in the manufacturing process of the p-n junction of the crystalline silicon solar cell core process, the conventional conventional closed-tube diffusion technology is generally used, and the diffusion device is shown in FIG. 1. However, the disadvantages of this technology are: 1) the consumption of the doping source is large; 2) the diffusion quality is easily affected by environmental changes such as exhaust air, and the equipment interferes with each other; 3) the pressure in the furnace mainly depends on the introduction of a large amount of nitrogen To maintain, the stability is poor, and gas is wasted [1,2].
The low-pressure diffusion technology can make up for many of the above deficiencies. The introduction of a vacuum environment greatly improves the diffusion uniformity and stability (within the range of up to 80 ~ 120Ω/port square resistance, the on-chip uniformity can be up to 4%, and the question and approval The uniformity can be up to 3%), the stable and uniform airflow in the vacuum chamber allows the placement of denser cells (the distance between the cells can be shortened to 2mm), which greatly improves the production capacity (single tube production capacity can reach 1000 or more) At the same time, thanks to the precise control and mixing of process gases in a vacuum environment, the amount of gas can be saved, and a cell with a higher square resistance can be made (a square resistance of up to 150Ω/port can be made). At present, low-pressure diffusion technology has been gradually recognized and applied in the industry [3, 4].
1. The principle and process of low-pressure diffusion technology
The working principle of the low-pressure diffusion furnace is that the reaction chamber is in a vacuum state through the action of a vacuum pump, and then the process gas is introduced. During the diffusion process, the chamber is maintained at a constant vacuum state, thereby ensuring that the concentration of phosphorus ions in the chamber remains unchanged. . As the free path of gas molecules increases under vacuum, the adhesion of gas molecules to the surface of the silicon wafer is also greatly improved, and a large amount of gas turbulence is avoided. At this time, the phosphorus ions can be evenly distributed on the surface of the silicon wafer, and the silicon wafer A uniform layer of impurity atoms is diffused on the surface to ensure the uniformity of the silicon wafer diffusion. Because the air flow in the vacuum chamber is more stable and uniform, allowing the placement of higher density silicon wafers, the distance between the silicon wafers can even be shortened to 2mm, which can double the diffusion capacity. Generally speaking, low-pressure diffusion technology is similar to the working mode of PECVD or LPCVD, running diffusion and doping processes [3, 4].
Figure 2 shows the core structure of the low-pressure diffusion furnace. The process gas enters from the intake pipe, and the low-pressure diffusion takes place in the quartz diffusion pipe. The exhaust gas is discharged from the exhaust pipe. The vacuum pump is responsible for creating vacuum conditions. The pump outlet is connected to the wind pressure of the acid discharge pipe.
Due to the introduction of the vacuum system, the low-pressure diffusion process is slightly different from the conventional diffusion process, as shown in Figure 3.
According to the principle of the low-pressure diffusion process, the low-pressure diffusion step adds the steps of vacuuming, leak detection and inflation; at the same time, the parameters that can be adjusted are increased, and the pump pressure, source bottle pressure and a series of unique adjustment characteristics of the vacuum atmosphere are added. When the process is operated under vacuum, the amount of N2 used is greatly reduced. Because of the above differences, the adjustment experience of the conventional diffusion process is no longer applicable to the low-pressure diffusion process, and it is necessary to re-find the effect of different diffusion parameters on the sheet resistance.
2. Experimental content of low pressure diffusion process
Commonly used parameter adjustments for low-pressure diffusion processes include diffusion temperature, source flow rate, diffusion time, source bottle pressure, and pump pressure. Among them, the adjustment of the diffusion temperature is the same as the conventional diffusion process. The adjustment of the diffusion time and the source flow rate is slightly different from the conventional diffusion process. The adjustment of the source bottle pressure and the pump pressure is a newly added adjustable parameter of low pressure diffusion. This article mainly studies the above parameters, and provides a certain idea and direction for the adjustment of low pressure diffusion process.
2.1 Study the effect of source flow on diffused square resistance
Source flow adjustment is the most commonly used method to adjust the diffusion block resistance. With other parameters unchanged, experiments were performed on the source flow of 700 sccm, 800 sccm, and 900 sccm to observe the effect on the diffusion block resistance.
It can be found from Fig. 4 that as the source flow rate decreases, the diffusion square resistance gradually increases. This is mainly because the source flow rate is reduced, the phosphorus element diffused on the silicon chip is reduced, the surface square resistance will become larger [5], so the square resistance will become larger. In Fig. 4, every time the source flow rate decreases by 30 sccm, the square resistance increases by 1 Ω/port, showing a linear relationship.
The source flow rate cannot be reduced indefinitely. When the source flow rate is reduced to below 700 sccm, the uniformity of the square resistance of the whole boat will become worse.
The block resistance has a certain gradient change from the furnace mouth to the furnace tail. The block resistance of the furnace mouth and the furnace tail differs by 15Ω/port, as shown in Table 1. This is because when the source flow rate is reduced to a certain amount, POCI3 cannot be evenly distributed in the furnace tube, which will result in a large concentration of POCI3 at the end of the furnace and a small amount of POCI3 at the furnace port. Therefore, there is a gradient of resistance from the furnace port to the end of the furnace. Happening. Special attention must be paid when adjusting the source flow.
In the actual adjustment process, it is recommended that the pure source flow rate change is ±50sccm, because after exceeding this empirical value, it is necessary to adjust the pressure of the source bottle and the amount of nitrogen to match the most changes in the source to ensure that the phosphorus source is on the surface of the silicon wafer. Evenly distributed.
2.2 Study the relationship between the resistance of different diffusion squares and the diffusion time
When the other parameters of the low-pressure diffusion process are unchanged, adjust the diffusion time so that the diffusion block resistance reaches 80Ω/port, 85Ω/port, 90Ω/port and 95Ω/port respectively. Observe the corresponding relationship between the diffusion block resistance and the diffusion time.
With the gradual increase of the square resistance, the diffusion time becomes shorter and shorter and the logarithmic relationship with the change of the square resistance, that is: the smaller the square resistance, the longer the adjustment time; the larger the square resistance, the longer the adjustment time less.
When adjusting the high square resistance, the time adjustment becomes less and less; that is, when the time change is small, the square resistance will have a certain change, and the stability of the high square resistance process will become worse, so adjusting the high square resistance cannot be blind To reduce the diffusion time, it needs to be adjusted together with other parameters.
2.3 Study the effect of the source bottle pressure on the diffusion square resistance
The source bottle pressure is one of the unique adjustable parameters of the low-pressure diffusion process. Studying the effect of the source bottle pressure on the resistance of the diffusion block is more meaningful for understanding the low-pressure diffusion process. When the vacuum pump pressure is 100 mbar, adjust the source bottle pressure to 770 mbar, 800 mbar, 830 mbar, and 860 mbar to perform the experiment, and observe the effect on the resistance of the diffusion square.
The pressure of the source bottle refers to the pressure at the outlet of the source bottle. The pressure of the source bottle controls the pressure at the outlet of the source bottle through an angle valve, thereby controlling the saturated vapor pressure in the source bottle. Increasing the pressure of the source bottle can reduce the source concentration, that is, diluting the phosphorus source; lowering the set value can increase the source concentration, that is, the total amount of source can be increased when the flow rate of the bubble carrier gas is unchanged [6].
The experiment found that with the increase of the pressure of the source bottle, the diffusion square resistance gradually increased. Since the pressure of the source bottle increases and the concentration of the source decreases, the square resistance becomes larger. Each time the pressure of the source bottle increases by 10 mbar, the square resistance increases by 1 Ω/port, showing a linear relationship. It can be seen from the data that the sensitivity of the source bottle pressure is better than the sensitivity of the source flow.
However, because the pressure of the source bottle affects the concentration of the source, it is not advisable to simply increase the pressure of the source bottle to greatly increase the square resistance. Due to the sharp reduction of the source concentration, the source in the reaction chamber cannot be uniformly attached to the surface of the silicon chip, which will cause the chip Uniformity becomes worse.
2.4 Study the effect of pump pressure on the resistance of the diffused square
Pump pressure is another adjustable process parameter unique to the low-pressure diffusion process, because the role of the vacuum pump determines the difference between low-pressure and conventional processes and is the core of the low-pressure process. When the source bottle pressure is 800 mbar, adjust the pump pressure to 55 mbar, 70 mbar, 85 mbar, and 100 mbar to conduct the experiment, and observe the effect on the diffusion square resistance.
The pump pressure is used to adjust the vacuum pressure in the reaction chamber. The vacuum pressure is generated by the collision of gas molecules. Different pump pressures represent different gas molecule concentrations. When the pump pressure is reduced, the movement speed of POCI3 in the furnace tube is accelerated. At the same time, it will be pumped away by the pump more, and the concentration is relatively low, resulting in the reduction of the phosphorus element left on the silicon wafer, which makes the square resistance increase: on the contrary, when When the pump pressure is increased, the square resistance will become smaller. When the pump pressure is small, the sealing performance of the vacuum system is higher, which may reduce the life of the seal and even explode the quartz tube, so the pump pressure cannot be reduced without limit [7 ].
实验可以发现，随着泵压力的增加，方块电阻变化呈乘幂变化，泵压力由55mbar增加到70mbar时，方块电阻减少10Ω／口；泵压力由85mbar增加到100mbar时，方块电阻仅仅减少3Ω／口。由此可以推算，当泵压力在100mbar以上时，对方块电阻的影响会减小。
本文分析了低压扩散较常规扩散的优势，论述了低压扩散的工作原理及可以调节的工艺参数。通过研究源流量、扩散时间、源瓶压力、泵压力等扩散工艺参数对扩散方块电阻的影响，找到这些参数与不同扩散方块电阻的对应关系，源瓶压力和源流量与方块电阻的变化呈线性关系，泵压力与方块电阻的变化呈乘幂关系，方块电阻与扩散时间的变化呈对数关系。了解低压扩散工艺的调试思路、途径及方法，可以更高效快速地开展低压扩散工艺的调试。
本文没有提及N2流量对低压扩散工艺的影响。在低压扩散工艺中，由于通过真空泵的作用，使反应腔室处于真空状态，N2的用量较常压工艺少很多，但N2用量更为精确，这体现在其对磷源浓度的影响和磷源在反应室中的停留时间。因此，N2对方块电阻也有一定的影响，而且会影响片内和片间方块电阻均匀性，今后可作为研究的方向。
参考文献
[1]Michael Quirkl．半导体制作技术[M]．北京：电子工业出版社．2009．
[2]安其霖，曹国深，刘祖名，等．太阳电池原理与工艺[M]．上海：上海科学技术出版社，1984．
[3]黄汉尧．半导体器件工艺原理[M]．北京：国防工业出版社，1980．78 107．
[4]刘恩科，朱秉升．半导体物理学[M]．北京：电子工业出版社．1980，69．
[5]Zhao J，Wang A，Green M A．24．5％efficiency silicon PERTcells on MCZ substrates and 24．7％efficiency PERL cells on FZsubstrates[J]．Progress in Photovoltaics，1999，40(10)：47l一474．
[6]Green M A．Third generation for photovoltaics：Ultrahighconversion efficiency at low cost[J]．Progress in PhotovoltaicsResearch and Applications，2001，9(2)：123—135．
[7]Zhao Jinhao．Recent advances of high·efficiency singlecrystalline silicon solar cells in processjngfechnoIogies andsubstrate materials[J]．Solar Energy Materials&Solar Cells．2004．(821：53—64）