In his role at Cascade Microtech, he frequently gets asked the question, “How do I know if my RF calibration is any good or not?” And it’s a question that should get asked. If your RF calibration is inaccurate then your measurements most certainly will be as well.

In this short presentation, he shows you ways to validate your calibration when using microwave probes and remove that doubt, ensuring that not only will your calibration be accurate but increasing your measurement accuracy in the process.

You can view the video at this link: http://ieeetv.ieee.org/player/html/viewer#microapps-techniques-for-validating-a-vector-network-analyzer-calibration-when-using-microwave-probes

1. How does the Tesla system handle such high currents?

Our break-through high-current probe has been developed to reduce probe and/or device destruction at high currents. The probe can support an amazing 10A of current in continuous mode and up to 100A of current in pulsed mode. This is no small feat considering that the probe tip is designed for pad sizes as small as 850um2.

Typically, when a high current is applied to a device the device will heat up. This device heating is often referred to as a self-heating effect. In order to reduce the amount of heating in a device, the test technician will often use a pulsed current measurement. Even in this case, at very high currents, on the order of 10A or greater, the device will experience some level of heating. This device heating causes a buildup of residual capacitance that can impact the measurement accuracy. In order to reduce device heating, the probe tip has been designed to minimize contact resistance at the wafer to probe interface. In addition, the probe also has a means of distributing the current over a multiple contact points that are joined with a relatively massive hunk of metal that is used to pull heat from the probe tip.

In addition, the probe features a replaceable tip design that can be interchanged as residues build at the tip, ensuring a low contact resistance at the tip.

2. Can I make a low-leakage measurement at high voltage?

The Tesla system also features a high-voltage probe that ensures operator safety and a high-performance electrical measurement path. The high-voltage probe provides the capability to make a coaxial measurement up to 3000V and triaxial measurements up to 1100V. In addition, the probe features a replaceable tip that can be easily interchanged as contaminates build at the tip affecting the accuracy of the measurement. This functionality makes it possible for device engineers to understand more of the characteristics of their device in the off state, enabling a premium price sort for devices with superior leakage handling capabilities.

3. What factors affect contact resistance?

There are two major factors to consider when looking at contact resistance for on-wafer power device measurements; first is the wafer to chuck contact resistance. In order to make an accurate Rds(on) measurement, it’s required to keep this value as low as possible. The Tesla system minimizes wafer to chuck contact resistance in two ways; first by providing a gold-plated chuck top surface and second through a superior vacuum pattern throughout the entire chuck surface that provides the optimal wafer hold-down even for today’s thinnest wafers. 

The second factor to consider is the probe to pad contact resistance. This value can be affected by pad material, pad thickness, probe tip shape, and probe scrub. The Tesla system has a multi-finger flat shape tip that creates only an insignificant amount of contact resistance when applied to typical aluminum pads; thus providing a more accurate data value.

If you use the Tesla system – or are considering it – and have questions, drop a comment below and we’ll address it in a future blog.

In the following presentation, we highlight field-proven, relevant test solutions based on a cryogenic wafer-probe station. Special attention is given to overcoming the calibration standards instability, contact repeatability and reliability issues caused by the extreme environmental conditions. We will present the solution that enabled characterizing of RF MEMS devices at cryogenic condition with the benchmarking level of measurement accuracy and confidence.

Cycle time and engineering cost reduction require accurate device models and their statistics. The requirement for accurate high frequency testing to 110 GHz combined with the need to probe aluminum pads with low contact resistance led us to develop the Infinity Probe® that offers both high-frequency performance and low, stable con­tact resistance on aluminum pads. As a result, the Infinity Probe has a number of advantages when used with devices that have aluminum – or gold – probe pads. But let’s explore why aluminum pad probing is challenging.

So why is aluminum pad probing such a challenge?

A thin layer of aluminum oxide (about 60 angstroms thick) immediately forms on a bare aluminum surface when exposed to air. This layer must be penetrated by deforming the relatively soft underlying aluminum, to directly contact aluminum metal with a probe tip. Aluminum oxide forms on any aluminum exposed by probing, so the electrical contact resistance is difficult to maintain for a long measurement or for many contact cycles without cleaning the alumi­num and aluminum oxide that accumulates on the probe tips.

Any measurement that is sensitive to a series resistance is affected by contact resistance variations. Such measurements include characterizing high-Q passive structures (such as inductors, capacitors, or resonators), and long characteriza­tion tests that require repeatable device contact for time periods beyond a few minutes (such as transistor S-parameter mapping of many bias conditions). Further implications are that measurements of the same structure are repeatable for very few contacts. When repeatedly probing a single transistor, contact resis­tance changes will modify the operating point so that each time a different case will be measured.

These problems lead to a lack of confidence due to the inconsistency in the data. The result of inconsistent data is poor correlation of data between measurements from differ­ent measurement systems.

There are numerous factors that can affect the RF probe contact resistance on alumi­num. These can be grouped into 3 categories:

1. Probe factors: tip contact area, force applied, tip metallurgy, scrub, and cleaning process.
2. Device factors: pad metallurgy, pad thickness, pad size, wafer cleanliness, surface contaminants, dielectric compliance, and pre-existing pad damage.
3. Environmental factors: temperature, mechanical vibrations, and probing system thermal stability.

The first category specifies the factors related to the RF probe:

• The contact area. This interacts with the probe force to set the pressure exerted on the pad. Probes for aluminum pads typically use an area of 100 to 400 μm2.
• The force on the probe tip and how this affects the scrub. Typical force per tip ranges from 5 to 10 grams for aluminum pads. Only a small horizontal tip motion (scrub) is necessary to break through the 60 angstroms of aluminum oxide.
• Tip material. Tungsten is the conventional material, but a non-oxidizing metal is better.
• Cleaning. Depending upon the pad metallurgy, aluminum and aluminum oxide must be periodically removed from the tips. Although, ideally, the abrasive cleaning removes the aluminum and aluminum oxide, it also typically changes the shape and size of the tungsten tip, causing additional changes in contact force and contact resistance.

Any type of contact resistance comparison testing requires careful consideration of all these factors. We’d love to hear more about challenges you encounter with probing, from metallurgy to pad size, so please share them in the comments section below.

We will repair all single and dual ACP on-wafer probes, FPC and Infinity Probes® at a fixed price per the table below. All single-signal ACP, FPC and Infinity probes (frequency < 110 GHz) received in quantities of 12 or less will be repaired within one week.

Multi-contact on-wafer probes including Unity™, Eye-Pass®, ACP-Q, DCP and WPH-900 probes will have a fixed repair price less than or equal to 50% of the standard new price (+ $100 for international shipping and handling).

NOTE: This program does not include our family of |Z| Probes® and ProbeWedges™.

Some details to note:

• Only probes purchased after January 1, 2008 qualify for repair (serial number must begin with an 8, 9, C or D)
• Repaired probes will be "as good as new" and meet all current quality and performance specifications; probes that do not meet these requirements will be replaced and the original probe serial number will increment as described in the Terms and Conditions*
• Cascade Microtech will pay shipping both ways and importation charges to the US; customer pays import fees on return such as VAT/GST
• FedEx will be the only carrier accepted
• All payment information (or Purchase Order) must be submitted prior to issuing a Return Material Authorization form (RMA)
• Maximum of 3 repairs per probe; probe serial numbers will increment with each repair (XXXX-1, XXXX-2 and XXXX-3)
• A complete list of Terms and Conditions are available on our website

We anticipate that this program will allow you to extend the life of your probe inventory and help manage the cost of probe replacement. If you’ve taken advantage of our probe repair program, please let us know if it’s working as planned, and if you have been satisfied with the results.

To initiate a repair, view additional details, or to see a current list of repair prices and Terms and Conditions, please visit our Probe Repair page on our website.

Developing a measurement environment immune from various noise issues, such as circuit operating noise, phase noise and random retention noise, has been a costly and time-intensive pursuit. The Edge flicker noise measurement solutions specifically address all the challenges involved with precision flicker noise measurement, ensuring a noise-immune test environment from 1 Hz to 40 MHz. The Edge solutions also enable a seamless transition from 1/f and RTS noise measurement mode to DC parameter extraction mode over a wide temperature range (-60°C to 300°C).