IV and CV Measurement Using the AgilentB1500A MFCMU and SCUU

Application Note B1500-3

IntroductionParametric characterizationrequires both current-voltage(IV) and capacitance-voltage(CV) measurements. Recentdevelopments in process tech-nology typically require accu-rate IV and CV measurementsbe made with a single pass ofthe devices on the wafer. TheAgilent B1500A SemiconductorDevice Analyzer supports single-pass IV and CV measurementswithin the same mainframeusing a new, single-slot multi-frequency capacitance measure-ment unit (MFCMU) and twoSMUs.

Performing both IV and CVmeasurements on a single probestation is not easy. SMU-basedIV measurements use triaxialconnectors, while CMU-basedCV measurements use BNCconnectors. Switching betweenthese two measurements can beconfusing and time-consumingand can often cause measure-ment errors. An example wouldbe where the measurementcables are manually switched,while keeping the probingneedle on the wafer. In this case,the electrical charge generated

by friction when the cables areswitched may damage the device.In addition to cable connectionissues, an error compensationparameter associated with thecapacitance measurements mustbe set correctly in order to ob-tain accurate results.

The B1500A’s SMU CMU UnifyUnit (SCUU) solves these mea-surement problems. The SCUU

Agilent B1500A Semiconductor Device Analyzer

enables accurate and effortlessswitching between IV and CVmeasurement without the com-plexity and expense of an ex-ternal switching matrix.

This application note illustrateshow you can easily configure anaccurate IV and CV measure-ment system using the B1500A.

Figure 1 shows a basic cableconfiguration diagram for makingCV measurements on a wafer.First, the four terminal pair(4TP) extension test cable isconnected to the measurementterminals of the B1500A’sMFCMU. The 4TP configurationis similar to a Kelvin connec-tion in IV measurements, and itcan minimize errors caused bymeasurement cable parasitics. Italso provides the best accuracywhen extending the test cable.

To obtain the best accuracy, the4TP extension should be routedas closely as possible to thedevice under test (DUT). Thenthe shielded two terminal (2T)cable or three terminal (3T)cable should be connected tothe probing manipulator. Thepath to the probing needle isbasically a 3T configuration.

Figure 2 illustrates the differ-ence between the shielded 2Tconfiguration and the 3T con-figuration. In the shielded 2Tconfiguration, the measurementcable guard shields are connec-ted together at the end of themeasurement cable. In the 3Tconfiguration, the guard shieldsare not connected together atthe end of the measurementcable. The shielded 2T configu-ration produces more stablemeasurements and betteraccuracy than the 3T configura-tion. This is because the resid-ual inductance in the shielded2T configuration is a fixed valuedetermined by the constructionof the cable.

Making Basic Capacitance Measurements with the B1500A

2

Figure 2. Differences between the shielded 2T and the 3T cable configurations.

Residual inductanceis a fixed value

Residual inductancevaries depending onthe arrangement ofthe cables.

Figure 1. Basic cable configuration for manipulator connection.

The residual inductance doesnot change, even if you movethe cables going to the DUT,thus providing stable andaccurate results.

The residual inductance in theshielded 2T configuration is fixedbecause an induced current flowsin the guard shield that is ofthe same magnitude but in theopposite direction as the currentflowing in the center conductor.This induced current helps toseal the magnetic flux created

3

There are three basic steps youmust take in order to get accu-rate capacitance measurements:first, minimize errors caused byparasitics; second, stabilize theerrors; and, third, effectivelycompensate for the errors. Fig-ure 3 illustrates that using ashielded 2T configuration pro -duces a relatively short path tothe tip of the probe needle. Theshorter the path to the probetip, the less chance there is forparasitics to cause errors (atypical error would be residualinductance in the measurementcable). By using a shielded 2Tconfiguration all the way to theprobe tips, you minimize thelength of the shielded 3T cable.This minimizes the length ofthe unstable portion of the mea-surement setup, and makes iteasier to compensate for mea-surement error using compensa-tion techniques.

by the measurement current,thereby keeping the residualcable inductance to a small andfixed value. In contrast, thereis no return current path gener-ated in the 3T configuration. Inthis configuration, residual in-ductance is determined by therandom arrangement of thecables. Since the configurationof the cables varies in an un-controlled fashion, it is difficultto keep the residual inductanceconstant and stable. Vibrationin the test environment changes

the residual inductance of themeasurement cable, making itunstable. This makes it difficultto compensate for the residualinductance of the measurementcables in the 3T configuration.This is sometimes referred to asthe "return path issue," whicharises when the guard shieldsat the end of cable in the 3Tconfiguration are not shortedtogether. You can solve thisproblem by connecting guardshields as shown in Figure 2.

Figure 3. The residual inductance of the positioners does not significantly impact measurement accuracy.

4

Considerations when Automating IV and CV Measurements

As previously described in thisapplication note, making asimple, accurate CV measure-ment is accomplished relativelyeasily. However, it is not so easyin an automated test environ-ment when you want to accu-rately measure both IV and CV.

Figure 4 shows an ideal preci-sion, low-current IV measure-ment setup where the triaxialcable from an SMU connects tothe input terminal of the waferprober manipulator. The mea-surement center conductor andguard shield are routed to theDUT via the shielded probeneedle.

A contrasting yet similar setupfor a precision CV measurementis shown in Figure 5. The cablerouting after the 4TP- to-shielded-2T conversion looks similar tothe IV setup, but there are twomajor differences. First, wheretriaxial cable is used in the IVmeasurement setup, coaxialcable is used in the CV setup.Second, there is a guard con-nection wire in between themeasurement cable and themanipulator needle in the CVsetup, where there is none inthe IV setup.

Although the cabling configura-tions for making separate CVand IV measurements, as shownin Figures 4 and 5, are similar,the differences pose a signifi -cant challenge in constructing asingle configuration that allowsyou to switch between IV andCV measurements in an auto -mated test environment. Thefollowing section in this appli-cation note shows how you canuse the B1500A to make bothIV and CV measurements usinga single cabling configuration,without degrading measurementaccuracy, and without anyspecial knowledge regardingmaking CV measurements.

Figure 4. Illustration of an ideal (non-Kelvin) on-wafer IV measurement setup.

Figure 5. Illustration of an ideal on-wafer CV measurement setup.

Differences inDC and CV

5

Figure 6 shows the basic cableconfiguration for the B1500Awhen it is used for switchingbetween IV and CV measure-ments. The B1500A’s SCUU canswitch between IV and CV mea-surements without degradingmeasurement accuracy, and both1.5 m and 3.0 m cables areavailable to enable the SCUU tomount on the wafer prober. Themeasurement cables from theSCUU are extended to thewafer prober manipulator usingtwo triaxial cables or two pairsof Kelvin (force and sense)triaxial cables that satisfy therequirements of both IV and CVmeasurements. For both IV andCV measurements, the centersignal line of the triaxial cableconnects to the center conduc-tor of the measurement needle,and the driven guard of thetriaxial cable connects to theouter shield of the measure-ment needle. Tying the outershield to the guard of themeasurement needle protectsagainst outside noise.

Just as is shown in Figure 5,the guard of the CH and CLmeasurement terminals of theprober manipulator are shortedtogether at the end of eachcoaxial measurement cable inorder to establish an accurateCV measurement system. TheB1500A’s guard switch unit(GSWU) is used to short theguard of the measurement cablewhile CV measurements areperformed. The GSWU switch

opens automatically during IVmeasurements to prevent poten-tial SMU damage, since other-wise the guards of the two SMUs(which are presumably at differ-ent potentials) would be shortedtogether.

Even if you implement a Kelvintriaxial connection between theSCUU and the prober manipu-lator, it will not affect the mea-surement accuracy if the cableis not too long. All you have todo is establish this simple con-nection system as well as collectand save the residual error com-ponents data using the open/short and (optional) load com-pensation routines built into theB1500A’s EasyEXPERT software.

Agilent B1500A IV/CV Solutions

The EasyEXPERT software han-dles all IV-CV switching anderror compensation and, bycontrolling the GWSU, alsohandles capacitance measure-ment current return pathissues. The EasyEXPERT soft -ware provides more than 100 IVand CV application tests. Yousimply select a CV algorithm,and push a button to beginmaking accurate CV measure-ments. If you use the SCUU, youcan extend the DC bias voltageof CV measurements to ±100 V,well beyond the inherent ±25 Vcapability of the MFCMU. In thecase of the SCUU, the ±100 Vof DC bias is automatically pro -vided by the SMUs connected tothe SCUU.

Figure 6: B1500A implementation of on-wafer CV measurement using the SCUU and GSWU.

6

Figure 7 shows how theB1500A’s MFCMU and HRSMUsare connected to the SCUUusing a 3 m SCUU cable adap-ter. The GSWU is connected tothe SCUU control terminal,which allows the GSWU switchto be controlled automatically.

Figure 8 shows the cablingfrom the SCUU and GSWU tothe wafer prober manipulator.The GSWU is connected to theguard of the wafer probermanipulator needle near thebase of the manipulator needleholder. This connection createsa return path when CV mea-surements are performed, thusensuring accurate and stablemeasurements.

Figure 7. B1500A’s MFCMU and HRSMUs connected to the SCUU using an SCUU cable adapter

Figure 8. SCUU and GSWU cabling to the wafer prober manipulator

Agilent B1500A IV/CV Solutions (continued)

7

The Agilent B1500A, in combi-nation with it’s MFCMU, SMUs,SCUU and GSWU, provides anaccurate and effortless systemfor switching between IV andCV measurements.

The B1500A’s EasyEXPERT soft-ware handles of all of the IV- CVswitching and error compensa-tion. You have only to select anIV or CV algorithm and push abutton in order to begin makingaccurate measurements.

Conclusion

For more information about Agilent and itsproducts, go to www.agilent.com.

For more information about Agilent Technologiessemiconductor test products, applications, andservices, visit our Website: www.agilent.com/go/semiconductor or call one of the centerslisted and ask to speak with a semiconductortest sales representative.

AmericasBrazil (11) 4197-3600Canada (French) 1 877 894-4414Canada (English) 1 800 447-8378Mexico 33 134-5841United States 1 800 447-8378

Asia/Asia PacificAustralia 1 800 629-485China 1 800 276-3059Hong Kong 852 2599 7889India 91/11 690-6156Japan 0120 421-345Malaysia 1 800 880-780New Zealand 0 800 738 378Philippines 1 800 1651-0135Singapore 1 800 276-3059South Korea 080 778-0011Taiwan 0 800 047-662Thailand 1 800 2758-5822

EuropeAustria (01) 25 125-7183Belgium (0) 2 404-9380Denmark 080301040Finland 20 547-9999France (0) 825 010710Germany (0) 18 05 24-63 34Greece 20 547-9999Ireland 016158393Italy 02 92 60 8333Luxembourg (0) 2 404-9340Netherlands (0) 20 547-9999Poland 20 547-9999Russia 20 547-9999Spain 91 631 3383Sweden 020 120-9975Switzerland (Italian) (0) 2 92 60 8484Switzerland (German) (0) 1 735-9300Switzerland (French) (0) 825 010 700United Kingdom (0) 7004 222-222

Middle EastIsrael 20 547-9999

Technical data subject to change without notice.

© Agilent Technologies, Inc. 2005Printed in USA August 1, 20055989-3608EN