Information about http://www.aecouncil.com/Papers/aec1.pdf
th ISTFA '93: The 19 International…
Tags: angeles california, body model, delco electronics, electronics industry, electrostatic discharge, failure analysis, human body, international symposium, kokomo indiana, machine model, model mm, pfaffenbach, phenomena, protection circuitry, robustness, semiconductor device, semiconductor devices, servais, voltage level, voltage levels,Pages: 7
Language: english
Created: Thursday, May 20, 1999 3:56:10 PM
Display cached document
Page 1
Page 2
Page 3
Page 4
Page 5
Page 6
Page 7
th
ISTFA '93: The 19 International Symposium for Testing & Failure Analysis, Los Angeles, California, USA/15-19 November 1993
An Investigation of Human Body Electrostatic Discharge
M.A. Kelly, G.E. Servais and T.V. Pfaffenbach
Delco Electronics
Kokomo, Indiana
Abstract Electrostatic discharge (ESD) events are
recognized as a significant contributor of early life
The electronics industry has recognized the failures and failures throughout the operating life of a
significance of Electrostatic Discharge (ESD) as a semiconductor device. Although present integrated
potential source of damage, especially to circuit designs include ESD protection circuitry, the
semiconductor devices, for some time. During that effectiveness of this protection must be determined
time, there has been an ongoing effort to develop a in a manner which will ensure its effectiveness in the
meaningful human body ESD pulse and equipment "real world" if the part is to meet the reliability
which is capable of repeatedly applying that pulse at requirements of the application.
various voltage levels to a semiconductor device. The ESD has been studied for some time, and there is
intent was to determine a part's ability to withstand an reasonable agreement on three (3) models for this
ESD pulse at a certain voltage level and use that phenomena: The Human Body Model (HBM),
information as an indicator of the part's robustness. Machine Model (MM), and Charged Device Model
Presently, available equipment is capable of applying (CDM). In this paper, we will be focusing on the HBM
an ESD pulse frequently described in specifications and some concerns we have about the model as
such as MIL-STD 883C as the human body pulse; but presently defined.
is this the right pulse? Recent technical papers have Under various conditions, the human body can be
raised some interesting questions about the ESD charged with electrical energy and transfer that charge
waveform and methods for capturing this waveform. to a semiconductor device through normal handling or
Specifications such as IEC 801-2 have also assembly operations. To evaluate the effectiveness of
contributed to the apparent confusion on ESD the protection circuitry in an integrated circuit, HBM
waveforms and, together, these sources of information ESD testing is performed. This HBM pulse is intended
were the catalyst that stimulated this investigation. to simulate the human body type ESD conditions the
part would experience during normal usage. The ESD
testing is also used to determine the immunity or
susceptibility level of a system or part to the HBM
TODAY'S ELECTRONICS INDUSTRY applications ESD event. Several different Human Body Model
place an increasing number of requirements upon (HBM) ESD simulation circuits and pulse waveforms
systems and component devices: semiconductor exist, including Military Standard MIL-STD 883C [1]
packaging and feature size is smaller, power (see Figure 1), International Electrotechnical
requirements and operating temperatures are higher, Commission (IEC) 801-2 [2] (see Figure 2), and
and reliability demands have increased significantly. others.
Designing for the elimination of early life device The basic objective of this paper is to explore the
failures is a key factor in meeting these reliability following question: Do the present test specifications
requirements. dealing with the human body ESD event define a
167
realistic ESD threshold or level of immunity, for a Limitations of Present ESD Methods
system or part in the real world? While the human
body ESD waveform has been a topic of research for A discrepancy appears to exist between reality,
many years, studies by Hyatt and Mellburg, "Bringing measured reality, and common practice as defined in
ESD Testing Into The 20th Century" [3]; Mellberg, some industry specifications. We feel a universally
Sanesi, and Hish, "Recent Developments In ESD accepted specification defining the actual human body
Waveform Evaluation" [4]; and Fisher, "A Severe waveform is not presently available due to various
Human ESD Model For Safety and High Reliability factors including:
System Qualification Testing" [5] have sparked our 1. The non-uniform conditions involved in
interest into the actual human body ESD event and the ESD environment.
waveform characteristics. 2. The unpredictable circumstances of the
ESD event.
3. The constant improvement in test
equipment used to study the ESD event.
Ip 4. Supplier community resistance to
Ipeak = 1.33 Amps
90%
adopting new standards that would
indicate some currently used protection
circuits are inadequate.
5. Lack of a standardized procedure for
capturing the ESD event. Some
procedures use measurement techniques
10% that are not capable of capturing the high
Tr < 10 ns Time (ns)
t frequency content or fast risetime of the
waveform.
Previous investigations into ESD testing have
resulted in two conflicting philosophies. One
Figure 1: Mil-Std 883 Human Body Waveform (2kV)
philosophy states, "the test procedure must look like a
human ESD spark...including all variability observed
in natural ESD phenomena" [3]. The second testing
Ip philosophy is to choose a representative waveform
Ipeak = 7.5 Amps from the range of likely ESD events and generate an
90%
instrumentation approach to ESD testing [3]. This
latter ESD testing philosophy employs test systems
designed to produce a consistent and repeatable ESD
waveform.
The difficulty with ESD test systems has been the
10% inability to deliver the relatively fast risetime
t associated with the surface charge stored on the
Tr < 1 ns Time (ns)
human body. Many test systems incorporate lumped
time constant circuitry and are plagued by parasitic
inductance, resistance, and capacitance of the various
components. These parasitics can greatly affect the
Figure 2: IEC 801-2 Human Body Waveform (2kV)
response of the ESD test system and therefore result
in invalid ESD event risetimes. The measured
To investigate the human body ESD event, a study risetimes are also limited by the capabilities of the
of the actual human body discharge was performed in measurement equipment used to capture the ESD
the laboratory. The intent of the investigation was to event waveform. When the MIL-STD 883C testing
gather a basic understanding of the HBM ESD event procedure was released in 1989, the risetime stated
and stimulate thought about the actual human body as less than 10 ns may have been accurate for the
discharge pulse and the possible effect on ESD type of equipment available for waveform verification.
immunity or susceptibility at the part level. Measurement equipment presently available is
capable of detecting and capturing ESD waveforms
with risetimes as fast as a few hundred picoseconds.
168
Electrostatic Charging 1000 to 2000 ohms is generally used. Recent
developments in the human body ESD model suggest
When two objects come in contact with each other, that the closet approximation to an actual human body
the triboelectric action between them can generate an ESD event may be the worst-case short circuit
electrical energy charge that initiates an ESD event. discharge waveform, documented by Fisher [5] and
The sudden release of generated charge in an object shown in figure 4, where the parameters are as
or person can produce extremely high voltages, follows: Rcharge = 1 M, Cbody = 60 to 300 pF,
currents, and electromagnetic fields that can result in Rbody = 150 to 1500 , Lbody = 0.5 to 2 H, Chand
malfunctioning, altering of device parameters, or even = 3 to 10 pF, Lhand = 0.05 to 0.2 H, and Rhand = 20
destruction of silicon junctions. In an ESD event, the to 200 . The equivalent circuit for the worst-case
human body can reportedly generate static charge scenario, the body-metallic object model, is found to
levels as high as 15,000 volts by simply walking produce a waveform almost identical to the actual
across a carpeted floor and 5,000 volts by walking human body discharge pulse captured during the
across a linoleum floor. The potential difference investigation we performed in the laboratory.
between a charged human body and an object
retaining an insignificant charge can range from a few
hundred volts to as high as 30,000 volts [6]. We tend Rcharge Rdischarge
to think these reported charge levels are exaggerated
1 Mohm 1500 ohm
due to the measurement errors reported by Ryser [7].
When a charged individual comes in contact with a
High Voltage
device or system, a transfer of the stored energy Pulse Cstray
DUT
100 pF socket
occurs to the device or through the device to ground. Generator
The typical ESD event has a fast, high current
peak followed by a lower, more slowly decaying
current pulse. The total energy in an ESD event can
be tens of millijoules with time constants measured in
picoseconds and several kilowatts of power [8]. With
this amount of energy available, it is quite evident how
a single ESD event can result in a device failure or Figure 3: Typical HBM equivalent circuit
possibly initiate a device weakness that can cause
failure with continued use.
Recent research on human body ESD events R
charge
R L L R
body body hand hand
shows that discharge pulses with fast risetimes, on the
order of 1 nanosecond or less, are the most disruptive
to the normal operation of electronic equipment [5].
High Voltage
Therefore, ESD test systems using a fast risetime Pulse C C DUT
body hand socket
pulse will more accurately simulate the human body Generator
discharge events frequently encountered.
Measurement of these parameters has been difficult
due primarily to the short time interval, large potential
differences, and the measurement bandwidth required
to capture both the amplitude and frequency
characteristics of the ESD event. These limitations Figure 4: Worst case, actual HBM equivalent circuit
may cloud the issues of ESD susceptibility levels and
environmental factors which may protect or damage Investigation Procedure
electronic devices.
The simplest human body ESD model is the series Before an investigation into the human body ESD
RLC circuit shown in figure 3 in which the R event could begin, the equipment used to capture the
corresponds to the body resistance, L is the discharge waveform must be understood. Most
corresponding body inductance, and C is the specifications require the use of an ESD target, or
capacitance of the body with respect to its current sensing transducer, to capture the human
surroundings. The body inductance is often body ESD waveform. This procedure involves
neglected, as in MIL-STD 883C, while a body connecting an ESD target to a digital signal analyzer
capacitance of 100 to 250 pF and body resistance of
169
or oscilloscope through low loss cables and having a Several individuals of various height, weight, and
charged "body" discharge by making physical contact gender were used as test subjects during the
with the target. investigation. One at a time, each "test subject" held
Due to the various ESD targets currently available, a metallic rod (charge/discharge probe) firmly in one
a greater understanding of each target and all hand and stood on the glass insulating plate. The test
corresponding characteristics was necessary. A study subject was then charged to a voltage potential by
by Mellberg, Sanesi, Nuebel and Hish [4] evaluated touching the metallic rod to a current limited high
the performance of several ESD targets including the voltage power supply. Once the individual was
Pellegrini, Electro-metrics, Mellberg-Hyatt, and saturated at a voltage potential, approximately 5 to 10
Reynolds-King versions. All targets were seconds, the metallic rod was removed from the
characterized by measuring insertion loss, impedance, power source and the charged subject was discharged
voltage standing wave ratio, and reflection (time into the target. Test subjects were asked to discharge
domain reflection). Based on the evaluation results, into the Pellegrini target using the metallic rod and
the Pellegrini type target was chosen. Another study again using their finger tip.
[9] recommends the use of a non-inductive current A typical 2000 volt (2 kV) discharge waveform
shunt which adds very little impedance into the ESD using the metallic rod is shown in Figure 5, while the
discharge path. The study also states the Pellegrini discharge waveform using a finger tip is shown in
probe introduces a front surface cavity that distorts the Figure 6.
local fields from a changed body prior to discharge.
The result of the distortion produces the dip seen on
the IEC 801-2 waveform (see figure 2).
Ip
Because the Pellegrini target was more readily Ipeak = 8.46 Amps
90%
available, it was chosen for waveform measurements Tr = 844 ps
in our investigation. The information presented in the
Hyatt study [9], however, should be taken into
consideration and researched further in order to
develop a consensus on the correct measurement
target to be used. 10%
The CT-1 current transducer, as specified in the t
MIL-STD 883C method, was also used to capture Time (ns)
ESD events. As will be shown later, the parasitic
inductance that is introduced by the CT-1 probe into
the ESD discharge path results in a measurement that Figure 5: HBM discharge waveform, 2 kV charge
indicates slower risetimes and dampened first peak metallic rod
amplitude characteristics.
In order to minimize errors with environmental
conditions, all testing was performed in a controlled
Ip
environment room which maintained temperature and
humidity at relatively constant levels of 23 +/- 4°C and
32 +/- 5% relative humidity. All human body ESD
events were captured using a 1 GHz real time Ipeak = 1.4 Amps
90% Tr = 727 ps
bandwidth digital signal analyzer with a 2
Gigasamples/second sampling rate and utilizing the
Pellegrini (1984) target as described in the IEC 801-2 10%
standard. The measurement equipment setup also t
Time (ns)
included a 36" x 18" x 3/8" (L x W x H) insulating
glass plate, high voltage power supply, ESD simulator
source, and several metallic rods to be used as
charge/discharge probes. The Pellegrini target was Figure 6: HBM discharge waveform, 2 kV charge
centered in a 1.5 meter square ground plane and finger tip
connected to the Digital Signal Analyzer through a 20
dB attenuator. The outstanding feature of these actual human
body ESD waveform measurements is the fast, high
amplitude initial peak followed by a secondary peak of
170
lower amplitude, significantly slower risetime, and the human body ESD event is not dependent upon the
longer dwell time. The observed first peak occurred in charging source.
800 to 1700 picoseconds, followed by a decay (20% of In order to examine the effects of higher voltage
first peak to zero) of approximately 100 nanoseconds. potentials, test subjects were charged to 8 kV and
The first peak amplitudes for a 2 kV potential were discharged into the Pellegrini target (see Figure 9).
measured and found to vary between 5 and 12 Amps, Although the amplitude and energy increased, the
depending on the individuals being charged. waveform remained essentially identical to the
Although the test subjects ranged from 125 to 325 previous discharge waveforms (see Figures 5, 7, and
lbs in total body weight and from 5'3" to 6'5" in height, 8). At this higher potential, however, a mild shock
no relationship between body size/shape and was felt by the individual when using the steel rod to
discharge peak amplitude was observed. Individuals discharge into the Pelligrini target.
that tended to be on either extreme of the height and
weight range resulted in the lowest discharge
waveform parameters (risetime and first peak Ip
amplitude). The human body ESD event is therefore 500 pF Capacitor (3.7 A & 1.12 ns)
90%
believed to be dependent upon skin-surface ESD Simulator (8.0 A & 983 ps)
resistance and/or body chemistry. To fully understand HV Power Supply (9.1 A & 818 ps)
the relationship between the HBM ESD event and the
characteristics of the human body, further
investigation is required.
The test subjects were then charged to a 2 kV 10%
potential and discharged using various metallic rods:
steel, brass, and a thin stainless steel screwdriver (see Time (ns)
t
Figure 7). The observed ESD waveforms were similar
to the waveform of Figure 5.
Figure 8: Discharge waveforms for a 2 kV charge,
various charging sources
Ip Steel rod (5.2 A & 852 ps)
90% Stainless Steel rod (8.2 A & 976 ps)
Ip
Brass rod (8.8 A & 796 ps) Ipeak = 31.2 Amps
90%
Tr = 1.2 ns
10%
t
Time (ns) 10%
t
Time (ns)
Figure 7: Discharge waveforms for a 2 kV charge,
various metallic rods Figure 9: HBM discharge waveform, 8 kV charge
The brass rod seemed to result in the highest first Capturing the ESD Event
peak amplitude, followed closely by the steel rod. The
different metallic material of each rod had little to no To investigate the effect of parasitic inductance
effect on the first peak risetime, therefore showing the on the measurement of the actual human body event,
risetime is dependent upon the individual or "body" the Pelligrini target was replaced with the CT-1 current
being discharged. transducer (1.5 GHz and 450 ohm) attached to the 50
Next, the test subjects were charged to a voltage ohm input of a 1 GHz preamplifier. Test subjects
potential using several sources: a high voltage power were charged to a 2 kV potential and discharged using
supply, an ESD simulator, and a 500 pF capacitor the CT-1 probe (see Figure 10). The use of the CT-1
(see Figure 8). Again, the discharge waveforms probe introduces approximately 90 nH of parasitic
resembled the typical waveform of Figure 5, showing inductance into the ESD discharge path. This
171
increased parasitic level resulted in the slower The waveform captured using the Pelligrini target
risetime, 15.55 ns as opposed to < 1 ns with the network, Figure 12, resembles the "shape" of the ESD
Pelligrini target, and dampened first peak amplitude, waveforms shown in Figures 5, 7, and 8 (although it
3.12 Amps as opposed to 7 Amps with the Pelligrini does not show the same first peak current levels).
target. Because the Pelligrini target network has a lower
The importance of the procedure used to capture parasitic inductance, the measured risetime and first
the ESD event cannot be overlooked. To highlight the peak amplitude (3.2 ns and 3.3 Amps respectively) far
effects of the measurement techniques used during exceed the waveform parameter limits defined in MIL-
this investigation and the resulting waveforms, several STD 833C.
commercially available ESD test systems were
characterized. These testers are used throughout the
electronics industry to evaluate the protection circuitry
Ip
of semiconductors or susceptibility level of a part to Ipeak = 1.31 Amps
an ESD event. The majority of ESD test systems Tr = 8.11 ns
implement the HBM waveform as defined in MIL-STD
883C. Prior to subjecting a part to the test system
90%
ESD event, a waveform verification procedure is
required. This is accomplished by capturing the ESD
discharge waveform created by the test system and 10%
verifying that the waveform parameters (risetime,
amplitude, etc.) are within the limits defined in the t
Time (ns)
ESD standard being implemented.
Figure 11: Test system ESD waveform (2 kV),
Ip
CT-1 probe
Ipeak = 3.12 Amps
90% Tr = 12 ns
Ip
Ipeak = 3.32 Amps
90%
Tr = 3.21 ns
10%
t
Time (ns)
10%
Figure 10: HBM discharge waveform (2 kV), t
Time (ns)
CT-1 probe
The question is, "What does the waveform actually Figure 12: Test system ESD waveform (2 kV),
look like?" The answer to that question greatly Pelligrini network
depends upon the method used to capture the ESD
waveform. Programming the ESD test system to a 2 While the same ESD test system produced both of
kV potential level, the resulting discharge waveform the ESD waveforms shown in Figures 11 and 12, the
was captured using the CT-1 probe (see Figure 11) and procedure used to capture those waveforms revealed
the Pelligrini target network (see Figure 12). two different ESD events. Due to these observed
The waveform captured using the CT-1 probe, differences, the question now becomes, "Which
Figure 11, reveals the dampening effect of the probe's waveform verification procedure is correct?" These
parasitic inductance. This particular measurement findings suggest the very real possibility that two (2)
procedure, which is identical to the waveform ESD test systems appearing to be equivalent using the
verification procedure defined in MIL-STD 883C, MIL-STD 883C waveform measurement procedure
resulted in the measured waveform parameters falling may be significantly different when measured with the
within the required limits of < 10 ns risetime and 1.33 Pellegrini target. Due to time constraints, these
+/- 10% first peak amplitude. questions could not be addressed in this study.
172
Further investigation in necessary, however, to References
establish an ESD specification and waveform
verification procedure that represents the "real world". 1. Mil-Std 883C Method 3015.7, Notice 8, 1989
Conclusion 2. International Electrotechnical Commission, IEC,
Standard 801-2, Second Edition, 1989
This work was not intended to be an all inclusive
study answering the questions which arise as you 3. H. Hyatt and H. Mellberg, "Bringing ESD Testing
become involved in ESD, but was intended to into the 20th Century", IEEE International
stimulate interest in additional work toward answering Symposium on EMC, 1982.
several basis questions. This evaluation has shown
that the true human body ESD waveform is 4. H. Mellberg, M. Sanesi, J. Nuebel, and A. Hish,
significantly different from that specified in MIL-STD "Recent Developments in ESD Waveform
883C, but is this difference important? Will it result in Evaluation", EOS/ESD Symposium Proceedings,
significantly different failure characteristics and 1991
thresholds in semiconductors tested with a fast
risetime, high energy waveform similar to the true 5. R. Fisher, "A Severe Human ESD Model for
HBM waveform? Safety and High Reliability System Qualification
This work also suggests that a significantly Testing", EOS/ESD Symposium Proceedings,
different ESD waveform may exist; one that is 1989
dependent on the measurement method. This could
be important when verifying whether two different test 6. W. Byrne, "Development of an Electrostatic
systems will result in identical ESD sensitivity level Discharge Model for Electronic Systems", IEEE
results. Additional work is required to established the International Symposium on EMC, 1982
appropriate test method and waveform verification
procedure. 7. H. Ryser, "The Relationship Between ESD Test
The testing performed with test subjects of widely Voltage and Personnel Charge Voltage",
varying body sizes and shapes demonstrated that EOS/ESD Symposium Proceedings, 1990
while size is not the significant factor in determining
the first peak power that will be observed in a human 8. B. Unger, "Electrostatic Discharge Failures of
body ESD event, skin conductivity, body chemistry, or Semiconductor Devices", IEEE International
some other factor may be of primary importance. Reliability Physics Symposium Proceedings, 1981
We feel this evaluation and many other studies
involving HBM ESD strongly suggest that some 9. H. Hyatt, "The Resistive Phase of an Air
present ESD test requirements should be Discharge and the Formation of Fast Risetime
revised/updated to reflect what appears to be a new ESD Pulses", EOS/ESD Symposium Proceedings,
and significantly more correct definition of the HBM 1992
ESD waveform.
173