A phase‐shift technique for high‐speed e‐beam testing with picosecond time resolution

Siemens Research Laboratories, Munich, Federal Republic of Germany
Scanning (Impact Factor: 1.89). 01/1989; 11(2):100 - 103. DOI: 10.1002/sca.4950110206


Electron beam sampling of high speed digital devices requires a high time resolution. At the same time, long-range phase shifting is necessary because signals in these circuits may have very long period lengths. In this article, a new phase-shift method is described which allows sampling of low repetition rate signals without any degradation of the time resolution. This long range phase shift is realized by an additional set of blanking plates or blanking capacitor, which, acting as a gate, selects one of a large number of electron pulses produced by a first blanking capacitor. This technique also allows fast switching between different phase angles. The phase-shift method was evaluated experimentally using the picosecond e-beam tester which was developed here. The time resolution of this tester has been optimized recently to allow for stroboscopic testing with a 7 ps pulse width at 20 mV/√Hz noise voltage and 0.5 μm spot size. This allows for the measuring of rise times down to 14 ps with an error below 10%. Phase shifts of 100 ns were realized without any degradation of this time resolution. Propagation delays of 3.5 ps could be resolved. Signal rise times of 40 ps, corresponding to 0.04% of the total delay could be easily measured.

2 Reads
  • [Show abstract] [Hide abstract]
    ABSTRACT: The aim of this paper is to point out the limits of the electron beam testability of MMIC given by the transit time effect (TTE). Theoretical investigations concerning the still open questions in connection with the electron beam testability of MMIC are thoroughly discussed. The calculations are based on the assumption of quasi-transverse electro-magnetic-waves which propagate as the fundamental mode on the MMIC typical waveguides, the microstrip lines and the coplanar waveguides. A criterion indicating the testability of a MMIC is defined. The microstrip line and the symmetric coplanar wave guide with different dimensions are examined. In addition, strategies for the reduction of the TTE, i.e. for the improvement of the electron beam testability of MMIC are proposed.
    Microelectronic Engineering 05/1990; 12(1-12):287-293. DOI:10.1016/0167-9317(90)90043-S · 1.20 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: The electron beam test (EBT) via the capacitive coupling voltage contrast (CCVC) is a promising method for a non-destructive testing of future integrated circuits (IC) with small structures and high speeds. In particular, it introduces the EBT into the field of IC production where only passivated IC exist and enables the EBT of multi-level-interconnection-IC. In a qualitative manner this method can be applied successfully for testing future IC. However, quantitative measurements suffer on errors which restrict the applicability of this method. These errors only permit the test of IC to line widths down to 3μm in the first interconnection level. Quantitative measurements at lower levels are even impossible. The errors can only be decreased by a design for EBT-ability.There is also a great demand for EBT of high speed devices, especially of monolithic microwave integrated circuits (MMIC). Each preparation of the IC-structure like the removal of passivation results in enormous functional changes. Therefore, an EBT of MMIC is only possible by the CCVC. Here the transit time effect (TTE) limits the time resolution. At a 5μm line a quantitative waveform measurement can be performed up to frequencies of 57GHz. Other measurement errors which are caused by microfield potential barriers and local field effects can be avoided by an adjustment of the operating point of the EBT-system.Although the EBT of passivated IC has been successfully performed there remains a problem, especially for long measurement times. The equilibrium of CCVC is not stable due to uncontrolled negative charging of the passivation. The reasons are still not clear.
    Microelectronic Engineering 05/1990; 12(1-4-12):325-340. DOI:10.1016/0167-9317(90)90046-V · 1.20 Impact Factor