Alfven Theory

Introduction

The Alfven | 100TM radio-frequency event detector application runs on our best-in-class VI technology platform. Unlike the Octiv range of VI probes, which measure rf current and voltage waveforms and frequency spectra, the Alfven | 100TM detects short lived excursions in the voltage and current amplitudes at the fundamental operating frequency.

RF processes are common in advanced manufacturing industries such as semiconductor, solar and thin film deposition. Plasma etching, physical vapour deposition and plasma enhanced chemical vapour deposition are prone to short lived events such as arcing, ignition phenomena and plasma instabilities. These events are typically in the range from 1 µs to a few 100 µs in duration. Significant damage to the features on the surface of the substrate being processed can be caused when these events occur. Thus, substrates may need to be scrapped at significant cost to the manufacturer.

The Alfven | 100TM design is coaxial in structure and has a characteristic impedance of 50 ohms. The input and output of the coaxial structure are terminated with standard RF connectors which are easily interchangeable. Therefore, the Alfven | 100TM can be fitted seamlessly in-line with the 50 ohm transmission cable used to carry the RF power from the generator to the matching network.

Generally, the events of interest originate on the process side of the matching network. Our technology allows the detection of the events on the generator side, through the matching network as depicted in figure 1.

Installation location of Alfven | 100<sup>TM</sup>
Figure 1: Schematic of the install location of the Alfven | 100TM event detector.

Theory of Operation

The Alfven | 100TM comprises a section of transmission line, voltage and current “pick-ups”, analogue signal conditioning circuits, ADCs, digital processing logic and a microcontroller as shown in figure 2. The RF signal going through the small section of transmission line is sensed using capacitive and inductive pick-ups to extract signals for the line voltage and current respectively. The pick-ups provide a low voltage signal compatible with the VI probe system electronics. These signals are digitized and processed within the logic circuit.

Alfven | 100<sup>TM</sup> sensor block diagram
Figure 2: Alfven | 100TM sensor block diagram.

The logic circuit uses a digital RMS detector to monitor the voltage and current magnitudes (at the fundamental frequency) continuously with 1 µs time resolution. An advanced algorithm searches for user defined ‘events’ which are unexpected in the RMS voltage and current data stream. If no events are detected, the algorithm averages the RMS voltage and current magnitudes and data is transmitted at a rate to 10 samples per second. If an event is detected, data is recorded with 1 µs time resolution for a period of 5 ms around that event.

The micro-controller retrieves the data from the Alfven logic. The micro-controller then applies scaling and calibration factors to the data to convert the processed digital values to real world voltage and current values. The controller will perform further averaging and report data at the requested rate.

Alfven | 100<sup>TM</sup> logic implementation.
Figure 3: Block diagram of the Alfven | 100TM logic implementation.

Events are detected and classified using the user defined criteria which specify the deviation from expected output in terms of:

  • magnitude of the deviation
  • direction of deviation (positive / negative)
  • duration of deviation

An event is defined as an unexpected change in signal that exceeds a specified threshold for a specified duration before returning to the expected level.

Alfven | 100<sup>TM</sup> event detection algorithm.
Figure 4: Schematic of an event as defined by the Alfven | 100TM detection algorithm.

Any signal change that remains above the threshold beyond the maximum event period is regarded as a change in signal rather than an event.

EVENT DETECTION

The Alfven provides a configurable event detection system, which includes a configurable noise filter, base signal threshold level, moving event detection threshold (based on a moving average and signal level) and configurable 3 x 3 event categorisation.

Alfven | 100<sup>TM</sup> illustration of detected event.
Figure 5: Illustration of a detected event.

CONFIGURATION PARAMETERS

There are a number of configuration parameters available to the user, these are listed below along with a definition of each parameter.

  • base_threshold
  • noise_filter_length
  • event_max_length_1
  • event_max_length_2
  • event_max_length_3
  • event_min_amplitude_1
  • event_min_amplitude_2
  • event_min_amplitude_3

BASE THRESHOLD LEVEL

Base threshold is used to configure the detection system floor level, signals below the base threshold will not be analysed for events.

NOISE FILTER LENGTH

The noise filter employs a configurable moving average filter, which is effectively a low pass filter. The length can be adjusted to smooth the analysed signal, removing false detections.

MAXIMUM EVENT PERIOD

This is the maximum length a detected event can be. Signal changes longer than this period are recognised as changes in signal level (ie step change). The maximum event period is equal to the maximum length in the configured event categories, i.e. event_max_length_3.

EVENT DETECTION THRESHOLDS

The event detection threshold is the minimum change in signal that will be recognised as an event. The threshold value changes with the signal, so that it is +/- X% around the moving average signal value, where X is the smallest detection threshold in the event category configuration, i.e. event_min_amplitude_1.

EVENT CATEGORIES

The events are categorised into 9 types, split into 3 event amplitudes and 3 event lengths. Events are categorised in amplitude where the events maximum percentage change from the average signal is used to define the event amplitude. Events are also categorised by length. The events length is defined as the length of time the event has been outside the event detection threshold i.e. event_min_amplitude_1.

< Length 1 < Length 2 < Length 3
> Amplitude 1 E11 E12 E13
> Amplitude 2 E21 E22 E23
> Amplitude 3 E31 E32 E33


In addition to these 9 event types it is possible to capture an event each time the RF power is switch on. This allows diagnostics of plasma strike events. When RF power switch-on events are enabled the sensor capture an event each time the RF voltage crosses the voltage defined by the BASE THRESHOLD LEVEL parameter. This type of event is labelled as E_ON in the Alfven software.

EXAMPLE CONFIGURATION

  • event_min_amplitude_1 = 5%
  • event_min_amplitude_2 = 10%
  • event_min_amplitude_3 = 20%
  • event_max_length_1 = 5 µs
  • event_max_length_2 = 50 µs
  • event_max_length_3 = 100 µs

An event with change of 15% and length of 45 µs will be categorised as E22.

  • event_min_amplitude_2 < 15% < event_min_amplitude_3, therefore fits Amplitude 2 category
  • event_max_length_1 < 45 µs < event_max_length_2, therefore fits Length 2 category

An event with change of 25% and length of 4 µs will be categorised as E31.
An event with change of 7% and length of 65 µs will be categorised as E13.

Any event with length greater than Length 3 will be recognised as a signal change, not an event.