A number of factors contribute to displayed value instability. Reading variation with respect to time is often the result of: 1) temperature changes that cause zero drift, 2) mechanical stress on sensor head assembly, 3) loose connections, 4) vertical position of sample in gap, and/or 5) sample characteristics.

This statement discusses three topics relevant to Conductance Monitor systems, and especially to instruments with the ten-fold sensitivity enhancement, (x10 instruments).

Contents: 1) reading variations due to environmental temperature changes, and physical or mechanical forces 2) techniques to assure proper performance of conductance monitor systems, and 3) cautions regarding calibration.

Slow numerical changes due to temperature change may be described as, "temperature drift." However, slowly changing physical forces, that may be incorrectly interpreted as temperature drift, can cause an operator to attempt an incorrect solution. Hence, careful analysis and distinction are required.

Reading variation of displayed value, as a result of temperature, is decreased by the appropriate design and use of integrated circuits with matched compensating silicon semiconductor junctions. Regardless, it should be noted that the temperature induced reading variation is primarily the result of thermal effects within the sensor head. In some instruments the sensor may cause 90% of the experienced drift. With our best systems the sensor head drift may be diminished to a level that matches the minimal, and predictable, drift value contributed by the electronic circuitry, hence, in the best situation the sensor and the electronics are equal and minimal contributors to temperature induced reading variation.

DELCOM customers exercise the freedom to purchase various combinations of available console units combined with numerous sensing head assemblies and options. Sensitivity enhanced, x10 instruments without temperature compensation have been specified, and purchased by DELCOM customers. This is not always a wise choice. The greater gain of tenfold affords a ten-fold improvement in precision, however the drift factor is also increased. Temperature compensation can decrease the drift by ninety percent resulting in greater reading stability of the last significant digit. Hence, temperature compensation is recommended as an option on all enhanced x10 instruments, especially remote types.

The electrical power delivered by the RTC module to the sensor head is measured in the low milliwatt range, hence, the electrical self-heating is non-measurable. Sensor head temperature changes are the result of the ambient environment. Embedded copper temperature stabilization lines are occasionally required to avoid the influence of ambient temperature changes. (a no-charge option on specified sensor heads) Users must not permit extreme temperature changes to occur. Most of our customers are controlling their stabilization coolant fluid to a small fraction of one centigrade degree.

Cable length in the vacuum chamber and inherent cable parasitic capacitance are critical controlling parameters. Both quantities should be minimized. In some cases we supply and install optional vacuum-tight boxes for containment of all critical components, primarily including the high frequency RF circuitry. These boxes are mounted on the sensing element assembly and reside in the vacuum chamber. Parasitic capacitance is then the function of predictable air molecules contained in the billet machined vacuum-tight boxes. Hence, digital signals leaving the Sensor Head/Vacuum Box assembly are insensitive to both cable length and cable capacitance changes. These systems operate very well!

Physical or mechanical concerns relevant to remote and some bench-top systems:

First, An operator of a Conductance Monitor should gain an appreciation of the instrument sensitivity by conducting the simple test of squeezing the sensor head assembly and noting the reading variation.

During operation the sensor head must not be subjected to any forces causing stresses that result in gap spacing changes. Problems will occur, if, during a vacuum chamber pump-down the deflection of the chamber walls cause forces to be applied to the sensor assembly. The best method for mounting a sensor assembly would be a knife edge at one end and a sliding surface at the other.

Secondly, it is necessary that users/customers notice and appreciate the importance of not disturbing the coaxial cable that leads from the RTC Module to the sensor head. Any repositioning of the cable can cause reading variation due to "internal cable capacitance changes" which result from the physical forces or changes in the path or position of the cable. For systems installed in vacuum chambers, pump-down also affects cable. Vacuum causes cables to expand and capacitance to decrease, thus, after pump-down, rezeroing of the instrument is required before the start of a coating operation.

Thirdly, It is of utmost importance, regardless of the quality of the mounting scheme at the sensor ends, that rigid fluid coolant lines not be directly connected to the sensor assembly. Flexible, woven cover, coolant lines should be used so as not to transfer force to the upper and lower bars of the frame.

Our instruments are calibrated using an electrically conductive material placed in the center of the gap. The center of a 0.100 inch gap is obviously at 0.050 inches. A non-conductive substance is used to support the sample at the level whereby the conductive surface of the film, or other conductive material, is centered at the above stated level. Note that some vacuum deposited materials change in sheet conductance value as measured in the chamber compared to the value measured at a later time at atmospheric pressure. This can be due to: 1) annealing, 2) effects of air etc. on the coating, such as oxidation, 3) scratches that are long enough and deep enough to cut through the conductive layer such that the circulating eddy current is prevented, and 4) other mechanical handling of the coated product.

Silicon calibration wafers can be used as calibration standards. If one attempts to four-point probe the wafer to arrive at a value, the result will not be the same as the "DELCOM value" labeled on a particular wafer. A difference of a number of percentage points will be noted. DELCOM wafers are indirectly calibrated from measurements taken from thin film samples measured with a four-bar system. At DELCOM, a stable and uniform thin film sample of size approximately two inches wide by eight inches long is placed in a fixture and Kelvin sensing is performed such that a center 2" by 2" square is evaluated and dedicated for use as a calibration standard in addition to equivalently qualified Si wafers.

Silicon wafers exhibit short term reading instability. This characteristic has been noted and documented by others including the NIST laboratory in Gaithersburg, Maryland, USA.. For some values of conductance, the time period to achieve a stable reading can be one minute, for some it can be as long as six minutes. The phenomena is not well understood, some maintain that all wafer testing and calibration should be done in a dark environment due to the release of photon stimulated carriers.

Conductance monitors have been our number one focus for more than fifteen years. DELCOM personnel have studied the attempts of other designers, have tested their methods and machines, and have obtained patents of those who think they possess better systems. There is no worldwide source for a better conductance monitor.