I-V Software

The first two sweep modes, V-I sweep and I-V sweep are continuous, linear-stair sweeps, and are similar in operation:

• Enter values for the sweep start and sweep stop
• Enter the number of points for the sweep, or alternately, click on the sweep step indicator to open a dialog box in which the desired step value can be entered.
• Enter the compliance setting, which is a limit value to prevent excessive current (or voltage) from being applied to the device under test.
• Select the sense mode:
  • 2-pt (local sense) uses basic two-wire connections to the device under test
  • 4-pt (remote sense) uses separate source and sense wires to the device under test. Also known as kelvin connections.
• The instrument's selected voltage and current ranges will be displayed. Unless auto-range is selected, sweeps will be performed using the fixed sense range shown.
• Press the "Execute" button to begin the sweep. The sweep progress will be shown on the SMU display, but the I-V software will be unresponsive while the sweep is running. The graph will be updated upon completion of the sweep. For particularly long sweeps, you may need to adjust the sweep timeout.

The third sweep mode, Diode sweep, is similar to the I-V sweep mode, except it performs a continuous, logarithmic-stair sweep. This is useful for measuring the exponential I-V behavior of diodes. Select the current range using the sliders. The polarity of the current can be positive or negative; however, the sweep always progresses from the smaller current magnitude to the larger current magnitude.

All sweep modes produce data sets that include voltage, current, time, and instrument-specific flags (e.g., compliance). The time values are reported by the SMU, which has an internal clock that is reset at the beginning of each sweep. By default, 24XX instruments report the measured (actual) source value, even if the instrument is in compliance. 23X-series instruments, however, report the commanded source value, which is not equal to the actual source value in cases of compliance.

Option: Auto-range

If selected, auto-range will instruct the instrument to automatically switch to the lowest available measurement range at each point of the sweep. This process is performed by the instrument, not the I-V software. Auto-range permits greater dynamic range in measurements. However, it can cause variations in sweep rates (high-sensitivity ranges have longer measurement delay), and it does not work well with certain devices or noisy configurations. Generally, auto-range is useful in improving measurement sensitivity. However, if a sweep returns puzzling results, or takes dramatically longer than intended, turn off auto-range as one of the first troubleshooting steps. Do not use auto-range in applications where constant dV/dt (or dI/dt) is desired during the sweep.

Comparison of diode V-I sweeps using fixed 10 mA range (left) and auto range with 100 mA compliance (right).

Unlike sweep measurements, DC (continuous) measurements operate indefinitely at a fixed voltage or current. Measurements are read continuously from the instrument, and are displayed on the graph in real-time.

To start operation in DC mode, first select the source mode (voltage or current) and the sense mode (2-pt or 4-pt). Then, enter the source value and compliance value, and select any other desired options. Press "Operate" to begin the measurement. The bias will be applied to the device under test, and measurements will be recorded continuously. Data will be plotted as a time series on the graph. Integrated charge and energy will be reported in the analysis results area (see power/charge integration).

Press "Operate" again to terminate measurements and remove the device bias.

When DC mode is operating, you can change the source current or voltage value as well as the compliance value. The changes will be reflected immediately on the instrument. Auto-ranging, integration time, averaging, and sweep delay can also be changed on the fly.

The time between measurements is determined by the software "sweep delay" setting, as well as instrument settings including the selected sensitivity range, integration period, and number of averages. Due to variations in communication and software execution speed, measurements do not occur at precisely timed intervals. Unlike sweep measurements, the time data values are not based on the instrument's internal timer; rather, they are based on a software timer. The accuracy of these time values is not specified. Thus, the time data might include a slight long-term drift that might vary from one computer to the next. Also, each data point's time value might differ slightly from the actual time of measurement due to variations in software execution order and speed. In practice, I have found the time values to be sufficiently accurate for all applications encountered to date.

Time-varying current in a forward-biased p-n junction diode, measured in a drafty room. The changes in current are primarily due to temperature fluctuations.

Option: Contact threshold

To assist with probing sensitive or fragile devices, this feature provides instant visual or audible feedback when contact has been made. To use: configure DC voltage source operation at a safe voltage such as 1V, with a reasonable current compliance such as 2 mA; then start DC-mode operation. With the probe(s) out of contact, the device current will be ~0 A. Set the contact threshold to 0.001 (A), and enable audible ("beep") and/or visual ("flash screen") contact indication. Then, approach the device with the probe(s). Contact feedback will occur once the measured device current reaches 1 mA.

It may be helpful to first touch the probes together to verify wiring and instrument setup.

The appropriate values for bias and contact threshold will depend on the I-V behavior of the device under test.

When sourcing current and measuring voltage, the contact threshold specifies a voltage below which contact occurs. Make sure to use a safe voltage compliance if such usage might expose the operator to live probe voltages--and never intentionally touch probes, conductors, or devices while the SMU is ON.

Beep means that the computer will play the default Windows notification sound, and that the SMU will beep (24XX series only).

Flash screen means that the software window background will turn bright red.

Screenshot showing DC-mode control options with contact indication.

The custom commands sweep allows you to manually specify commands to perform sweep types that are otherwise not supported in this software. Three text fields are provided for commands, the sweep config command, the on/trigger/read command, and the off command. These commands are combined with advanced instrument setting commands. When a custom sweep is executed, the following commands are sent to the instrument:

1. RESET command (J0XO0X for 23X; *RST for 24XX)
2. Adv. init commands, if enabled
3. Custom sweep config commands
4. Adv. pre-sweep commands, if enabled
5. Custom on/trigger/read commands

At this point, the data is read from the instrument. Optionally, the software will wait for GPIB SRQ before initiating the read. In general, 23X-series instruments should use the wait-for-SRQ read method, whereas 24XX-series instruments can be configured either way.

6. Custom off command
7. Adv. post-sweep commands, if enabled

In this mode, except for the initial RESET command, no other commands are automatically generated by the software. So, none of the instrument timing settings from the timing and averages tab will be transmitted to the instrument. To assist in verifying (or debugging) a custom command sweep, turn on the "debug" option under advanced instrument settings.

After the sweep data has been returned, the software will parse the data and display it on the graph. The software can only process ASCII-formatted data, so the custom commands must instruct the instrument to return data in ASCII format. The preferred format commands are:

• 23X-series: G13,1,2X
• 24XX-series: :FORM ASC;:FORM:ELEM VOLT,CURR,TIME,STAT

There are several ways to export data from the I-V software:

• Save data. This button will open a file dialog to save the I-V data. The file format is tab-separated ASCII (suitable for opening in a spreadsheet program such as Excel). Saved files also contain additional segments including the analysis settings & results, and the data description field (if present).
• Copy data. This button copies the data to the clipboard in tab-separated ASCII format, suitable for pasting into most data processing or spreadsheet software. If selected, the analysis results (if displayed) are also included.
• Copy graph (as bitmap). To copy the graph as an image, exactly as it is displayed on screen, right-click on the graph area and select "Copy data." The graph will be copied as a bitmap image.
• Export graph (as image or data table). To export the graph in a different image format, or to directly export the data plotted, right-click on the graph area and select the "Export" submenu. With this approach it is possible to export the graph in vector graphics format (emf or eps).

To load prior trace data, use the "Load Data" button. The data file format much match that of files saved by this software, i.e., tab-separated ASCII with appropriate headers and/or metadata. If trying to import data from other sources, open a previously saved file in a text editor to view the file format.

The software provides numerous options to plot the measured data. Controls for zooming, panning, lin/log scale, and data selection are provided. The graph mode selector provides the following options:

• Current vs. Voltage
• Voltage vs. Current
• Current vs. Time
• Voltage vs. Time
• Power vs. Time (calculated for each data point as P=IV)
• Resistance vs. Time (calculated for each data point as R=V/I)

Illustration of graph controls.

A single (V,I) data point conveys the resistance (R=V/I) of a resistor, but it is often of interest to measure the average resistance over a number of bias values, or to measure the incremental (small-signal) resistance (R=dV/dI) of non-linear semiconductor devices such as diodes (series resistance) and transistors (output resistance). The resistance fit performs a linear fit to a selected region of I-V data, and is useful for these applications. Drag the graph cursors to indicate the start and stop of the fit range. The default type of fit is a linear regression fit.

Options:

• Hold shift while pressing the resistance analysis button to perform a simple, 2-point fit to the two cursors (instead of a regression fit).
• Double-click the resistance fit button to automatically select the fit range. The software will select the largest span of data points for which valid (i.e., non-compliance) measurements are present. This is generally useful for linear resistive devices, and is especially useful when using 23X-series instruments, which lack source read-back capability.

The 4-point probe is the standard instrument for quickly measuring wafer resistivity, or checking the sheet resistance of an epi or diffused layer. The probe comprises a Wenner electrode array with four colinear equally spaced electrodes. This configuration is illustrated below, along with the relevant equations. A brief discussion of the theory can be found here, or in any microelectronics fabrication textbook or website. However, note that terminology and symbol usage vary between authors; make sure to understand the purpose of the lateral correction factor C as used in this software.

4-point probe illustration and equations.

This software uses the general equations for sheet resistance and resistivity shown above (labeled "otherwise"). This resistivity equation already includes the first-order wafer thickness correction factor. The parameter C is a lateral correction factor, and comes into play when the lateral extent of the conductor (D) does not satisfy D >> s. This occurs with small wafers, or with small diffusion test areas on a process wafer. The value of C has been analyzed extensively for various test geometries (the "Haldor-Topsoe geometric factors"); as of this writing, two online resources can be found here and here.

Most semiconductors have well-established relationships between resistivity and dopant concentrations; as of this writing, a useful online calculator for silicon can be found here.

Options/tips:

• Hold shift while pressing the 4-probe analysis button to perform a simple, 2-point fit to the two cursors (instead of a regression fit).
• Double-click the 4-probe fit button to automatically select the fit range. The software will select the largest span of data points for which valid (i.e., non-compliance) measurements are present. This is particularly useful for 4-probe measurements with 23X-series instruments.
• When first finding the appropriate voltage/current sweep settings, it is useful to make use of the F3/F4, F9/F10, F1, and F2 function keys.

The solar cell analysis automatically extracts the open-circuit voltage, short-circuit current, fill factor, and other parameters related to photovoltaic behavior. Efficiency is calculated based on the provided values for the device area and incident power density. The analysis works regardless of the polarity of the photovoltaic device, i.e., results are extracted from either quadrant IV ("conventional" polarity) or quadrant II ("reverse" polarity).

Options/tips:

• Double-click the solar cell analysis button to automatically scale the graph on the region of photovoltaic power generation. A graph cursor will indicate the point of maximum power.
• The software can try to detect suspicious PV analysis results, by looking for non-monotonic fluctuations in photocurrent. The "flicker" parameter sets this threshold.
• The analysis will also extract the local slope (resistance) of the I-V curve at the open-circuit and short-circuit points. These values can be useful in identifying major problems with shunt and series resistance; however, unless the I-V data is "clean and smooth", these extracted values tend to reflect measurement artifacts. The "slope neighborhood" field sets the number of points over which each slope is calculated.

The diode analysis fits the exponential current-voltage behavior of a semiconductor p-n or Schottky junction diode. The result is a linear fit on a semilog plot, allowing extraction of diode parameters including ideality factor n and saturation current I₀. To use, position the graph cursors to select the region of exponential I-V behavior, i.e., the knee of the I-V curve. Then press the diode analysis button and view the fit on a semilog plot.

Options/tips:

• Normally, ideality factor is extracted assuming a junction temperature of 300K. If the device is at a different temperature, press shift while pressing the diode analysis button. This will bring up a dialog box to input the device temperature.

When performing DC measurements, or viewing I-V data plotted in "Current vs. time" or "Voltage vs. time" modes, the software can integrate the power and charge dissipated (or generated) by the device under test. This occurs automatically during DC measurements. To display integrated power/charge after a measurement is complete, switch to "Current vs. Time" or "Voltage vs. time" graph modes, then drag the cursors to define the start and stop of the integration period. Results will be displayed automatically.

As discussed in the description of DC-mode measurements, software timing issues might introduce slight errors in total power and charge calculations. These errors should be negligible in all but the most demanding applications.

Sometimes, the integrated power or charge values will report 'NaN' (not a number). This means that the selected time interval contains invalid measurement points, or measurement points from which power (or charge) cannot be determined. The most frequent cause is that the instrument reported compliance for one or more measurements. The 23X-series source meters lack source-readback capability, thus when compliance occurs, the actual power delivered to the device cannot be determined.