自动化蚀刻控制下的可重复性

摘要：以氟化氢为基础的溶液被广泛地用在扩散前清洗工艺中的二氧化硅清洗与蚀刻中。为降低不同批次之间工艺效果的差别，也就是增加批次与批次之间工艺效果的可重复性，更好地控制氧化物地腐蚀显得尤为重要。自动化的蚀刻控制体系弥补了在线的化学品浓度与反应温度的偏差而达到了的蚀刻目标。

关键词：可重复性；清洗；腐蚀
1 Introduction

Spray processors are dry-in/dry-out wet cleaning systems capable of being used at virtually all wet cleaning steps for both FEOL and BEOL applications. Batch spray processors are typically configured with up to eight chemical inputs, in addition to DI water, and up to two recirculation systems. Chemicals can be dispensed inpidually or mixed with each other and/or DI water. Once mixed the chemicals are dispensed onto the wafers in a process chamber, as shown in Figure 1. This mixing flexibility is used to generate a wide range of chemical mixtures that can be dispensed at ambient temperature or heated with an infrared chemical heater.

Spray processors have a number of advantages.They typically have a smaller footprint since all processing is performed in a single chamber. All chemical that is dispensed onto a wafer surface is either fresh or freshly filtered. Since the wafers are rotated during the cleaning process, centrifugal force enhances particle removal efficiency and overall rinse efficiency. Enhanced rinse efficiency lowers overall DI water consumables when compared to comparable immersion processes. While recirculation is available, for many applications it is not neededdue to the highly efficient chemical usage and low cost of consumption. The single-pass processing, made affordable by spray process, eliminates the risk of cross-contamination.

2 Oxide etching

Due to fluctuation in incoming DI water temperature, exothermic reaction of DI and HF mixing and different tool idle time, there can be significant temperature difference during etch process and among different batches. Figure 2 shows the actual process temperature from different runs. A temperature difference of 2℃ was observed, and it caused a large etch rate difference, as shown in Figure 3. Although the average etch rate was close to the target, the run-to-run etch rate repeatability deviation was about 6.5%, which was too large for requirements of the pre-diffusion clean process.

In order to reduce the etch rate variation and improve run-to-run repeatability, etch time should be controlled to compensate for temperature and concentration variation. Automated etch control was developed based on these concepts and is described in the following paragraphs.

3 Automated etch control

The etch control option automatically controls the length of etch steps by measuring and adjusting for concentration of the etchant and wafer temperature. The etch control process is defined by a base etch rate which assumes a constant temperature (base temperature) and a constant chemicalconcentration (base concentration).

Because variations in temperature and concentration affect the etch rate, etch control software examines real-time temperature and concentration and varies the time of the etch accordingly. In this manner, the software compensates for process variations and provides an accurate etch. This method utilizes the higher accuracy of chamber temperature measurements and flow stream concentration calculations with scaled integer arithmetic instead of the less accurate closed-loop approach. Each recipe that uses etch control must contain etch parameters. One parameter defines a base etch rate. A second parameter names the temperature at which the base etch rate is valid. The remaining parameter specifies the thickness of material to be removed in the process. Etch parameters are defined with a syntax similar to the syntax used for setting flow rates. The plumbing configuration for etch control is shown in Fig. 4. The flow measurements for diluent and etchant (required to calculate% concentration) are obtained by flow pickup D and flow pickup E respectively. The etchant and diluent are combined in the chemical manifold, and they flow as a mixture to the spraypost.

Temperature is measured by a low mass RTD located in the side of the process chamber. A second low-mass RTD is compared with the primary RTD to detect a defective temperature measurement. If both etchant and diluent are flowing during the etch step, the software will adjust the estimated etch rate based on the observed variations of the flow of the etchant and diluent.

4 Etch control implemenation

One parameter of etch control is temperature, and accurate temperature measurement is essential to keeping the etch control model valid. Temperature sensor is encased to protect against contamination. Figure 5 is the temperature sensor data between encased and non-encased side bowl temp probe. It was found that the temperature sensor was very accurate, and encasement of temperature sensor did not impact the response time.

Another parameter of etch control is concentration. The etch control model is effective for chemical concentration variations within 20%, where the etch rate is thought to be approximately linear with the concentration, and is compensated accordingly. The etch control software can compensate for HF flow rate variations of 20% or less. Figure 6 is an example of HF flow variations that can be compensated by etch control.

5 Automated etch control results

To overcome the etch rate variations shown in Figure 3, automated etch control option was applied to oxide etch with 100:1 HF solution in FSI MERCURY spray processing system, and the resulting oxide etch rate is shown in Figure 7. Better uniformity and less variation among different batches were obtained. For the same oxide loss target, as compared to the results without etch control shown in Figure 3, the standard deviation was reduced from 6.76A to 2.10A, and the repeatability deviation was dropped from 6.45% to 1.99%.