Electrical: Current-Voltage (I – V) Probe

In a plasma CVD process, the knowledge of the applied power and the understanding of the state of the plasma is important. Normally, the reflected power, recorded by a power meter is subtracted from the forwarded power to estimate the power applied to the plasma. However, the real power coupled to the plasma can be much less due to loss of power in the cables and in the matching network; the loss increases with increase in plasma frequency. Thus, while for St RF PECVD this loss may not be so severe, at VHF conditions it can be substantial. In a poorly designed VHF reactor and cable lines, the plasma may not even get ignited due to a substantial loss of power. For this reason, it is necessary to have a V -1 probe attached to the RF/VHF power feeding, between the matchbox and the reactor to estimate the real-time voltage |V| and current |I | and the phase difference between them, ф. The applied power is given by |V|*|I|*cos ф. The reactance X, resistance R and impedance Z are calculated from the relation,

V V

X = sinф, R = cosф, and |Z| = |X|2 – |R|2 (7.1)

It should be noted that the total impedance is a superimposition of chamber impedance and plasma impedance. The phase difference ф between current and voltage for a purely capacitively coupled plasma is ideally 90°. However, due to propagation delays in the probes [46] there might be a phase error of both voltage and current and the net effect is that the phase of the impedance (current with respect to voltage) is slightly different from the expected 90°. This phase error needs to be taken into account to calibrate the phase at any particular frequency. With this calibration, it is possible to record the phase change due to change of plasma conditions, such as in the case of high hydrogen dilution or of the so-called a to y ’ transition, which will be illustrated in Section 7.4.7.2. The latter type of plasma phase change condition is accompanied by a change of impedance from purely capacitive to resistive, manifested as reduction of phase. Figure 7.1 shows the change of phase with the applied power for low deposition rate and high deposition rate of nc-Si while the applied power is determined and controlled by the V -1 probe. A substantial change from a capacitive to resistive plasma is observed for the high deposition rate

83 82 – 81 80 – 79 – 78 – 77 76 – 75 – 74 – 73 72

Подпись: о a) w CB -C CL Подпись: H2/SiH4Подпись: Figure 7.1 Dependence of the phase of the plasma on hydrogen dilution at two applied power conditions [54].image2250

case. The Z-Scan RF probe and MKS RF impedance analyzer are excellent V -1 probes that may be operated at single or multiple frequencies.

A critical problem in plasma CVD processes is the detection of plasma transients, which could cause a significant change in the deposition conditions. The diagnosis of a transient plasma parameter signal is possible only when the response times of the detection process are smaller than that of the plasma transient. Many of the stan­dard diagnosis tools of the plasma such as DC bias and the output of the V -1 probe can follow the plasma transients, which are in the seconds range, such as transients related to gas compositions or a dusty regime (see Section 7.4.7.2), but are much slower than most of the timescales of electron impact dissociations and the concomitant tran­sient plasmas in the subsecond range. However, there are tools sensitive enough to follow these fast plasma transients, as shown below. The electrical parameters of the plasma can be detected by a fast oscilloscope (see Figure 7.2) because the plasma turn-on transients are in the order of microseconds. The detailed nature of a pulsed plasma can be seen in Figure 7.2, where an initial rise period is followed by an over­shoot of the voltage and then decay to a stabilizing voltage before sharp decay to the off period.