The health hazards of depleted uranium munitions

ANNEXE E Reviews covering characteristics of DU aerosols

M R Bailey (NRPB)

(Note that in this annexe our comments are in this format: ie, in italics between brackets, to
distinguish them from comments in the original reports.)

As a starting point in this assessment, a number of reviews and assessments were consulted
principally to obtain an overview of the likely scenarios, and to obtain references to ‘source’
documents, ie first hand reports of relevant observations, measurements or calculations. Relevant
aspects are summarised in the following subsections.

E1 AMCCOM (1990)

US Army Armament, Munitions and Chemical Command (AMCCOM) Task Group Report. Kinetic
Energy Penetrators: Environmental and Health Considerations (Abridged) Kinetic Energy Penetrator
Long Term Strategy Study, Picatinny Arsenal, NJ. (A copy of this report, entitled ‘Appendix D’ was
supplied by R Brown, DRPS, but without a reference list, which was still being sought. It is probably
‘Danesi M E 1990’ cited in AEPI 1995, although the name Danesi was not found in it.)

The introduction notes that the US Army was formulating a strategy for future kinetic energy (KE)
penetrating materials. The objectives were to perform a preliminary ass essment of the
environmental and health issues associated with DU and tungsten penetrator manufacturing,
testing and recycling facilities. (This provides a useful comparison between DU and tungsten, the
obvious alternative material for KE penetrators.) It included some requirements for decontamination
of equipment, and addressed combat issues.

As background (volume 1 page 1-2) it notes that DU penetrators had been produced since 1979,
replacing tungsten penetrators, which were used previously. Due to the radioactive and chemically
reactive nature of DU, concerns had arisen regarding environmental and health issues throughout
the life cycle of the ammunition. Concerns at manufacturing and testing facilities centred on
occupational and public exposures to DU along with potential contamination of the environment.

E1.1 Testing of penetrators

Testing operations are described in volume 1 pages 3-2 – 3-3 and volume 2 pages 2-2 – 2-4. It
notes that hard target testing involves firing DU penetrators into armour plate to demonstrate
penetration. On impact, the penetrator fragments, and portions of it burn forming uranium oxides
(mainly UO 2 and U 3O8). The degree of fragmentation and size of particles generated depend on
firing conditions and target characteristics. Previously some hard target testing of both DU and
tungsten penetrators was done in the open air. For the last several years, all DU penetrator hard
target testing within DoD had been done in containment facilities. A new enclosure was being
constructed on the Ford’s Farm range of Aberdeen Proving Ground, Maryland. This enclosure,
identified as the Depleted Uranium Containment Fixture (the ‘Superbox’), was scheduled to
become operational in mid-1990. It would permit the environmentally safe and effective testing of
DU materials, including firing DU penetrators into full sized, fully loaded armoured vehicle targets.
Upon impact with the armour plate some KE penetrator fragments penetrate the armour and some

1
are reflected from it. Both types are of wide size range, from large pieces of the order of
centimetres down to micron size. Most of the fragments spontaneously ignite due to the
pyrophoric nature of uranium and the extreme temperatures generated on impact (3037° –
3093°C) over the range of impact velocities (4010 – 5560 feet per second). Because of the high
density of uranium and its oxides, the penetrator fragments and particles rapidly settle in the target
area.

Soft target testing involves firing penetrators through a target, usually canvas stretched between
poles located either 1000 or 4000 metres downrange, to measure flight accuracy. After passing
through the target, the penetrator impacts the ground and is either stopped by a soil ‘berm’ or
skips along the ground until it comes to rest. At the point where the penetrators impact the ground
a trench develops, which can be as much as 600 m long by 10 m wide by 1 m deep. Fragmentation
of the penetrators during soft target testing depends on the type of soil and obstructions. Severe
fragmentation can occur where soils are hard and rocky and where there are obstructions such as
trees. Conversely, where the soils are very soft and where there are no obstructions, very little
fragmentation occurs. (This is relevant to dispersal in the battlefield environment when targets are
missed.)

E1.2 Health effects

With regard to health effects (pp 4-7 – 4.14) it notes that “Both DU and tungsten present the
potential for deleterious health effects”. However it was considered that the use of both in KE
penetrators could me et guidelines set by regulators at which the materials could be safely used.
Further research into both was needed, but “Uranium, in contrast to tungsten, has been extensively
studied due to its role in the nuclear fuel cycle”.

E1.3 Combat issues

Combat issues are described in volume 1 pp 4-5, volume 2 appendix D pp D-10 – D-12 and pp D-
18 – D-20 and appendix F. As this is mainly an assessment it is included in annexe F, section F1.

E2 AEPI (1995)

Health and Environmental Consequences of Depleted Uranium use in the U.S. Army: Technical
Report. (US Army Environmental Policy Institute)

E2.1 Introduction
In response to a Congressional request, the AEPI conducted a study to determine:
• the health and environmental consequences of using DU on the battlefield;
• remediation technologies that exist or might be developed to clean up DU contamination;
• ways to reduce DU toxicity;
• how to protect the environment from the long-term consequences of DU use.

AEPI based its study on a comprehensive review of the environmental safety and health literature
available in November 1993. The authors reviewed the four previous major DU studies:
• Medical and Environmental Evaluation of Depleted Uranium (JTCG/ME 1974);
• A Hazard Evaluation of the Use of Depleted Uranium Penetrators (Pierre Committee 1979);
• Comparison of DU and Tungsten for use as Kinetic Energy Penetrators (NMAB 1979);
• Kinetic Energy Penetrator Long Term Strategy Study (AMCCOM 1990, cited in AEPI 1995 as
Danesi 1990).

2
In summary (section 5.4) it notes that “The JTCG/ME study concluded that the overall use of DU
penetrators would have no significant medical or environmental impact but that, depending on
local conditions, an uncontrolled release of DU could have a significant impact. Further, it
concluded that implementing regulatory requirements of NRC, DOT and OSHA would effectively
control DU hazards during peacetime. It also concluded that fires or accidents involving DU
munitions and the use of DU munitions during combat could cause locally significant
internalization. The three later studies concurred with these conclusions”. (Thus the earliest (1974)
assessment seems broadly consistent with most later ones: generally exposures and risks should be
low, but in in some circumstances high intakes can occur. The most recent of these four previous
studies is addressed in section E1. We did not obtain the others.) It is noted that these reports
provided a relatively complete picture, but the US Army had not pursued many of the health and
environment-related studies recommended by them. To extend and update the information, the
study held numerous discussions with personnel involved with DU munitions.

Part 1 (chapters 1 – 4) discusses DU and its use by the US Army; part 2 (chapters 5 – 8) covers
responses to the tasks assigned by Congress.

E2.2 Properties, characteristics and life cycle of uranium used by the army
Chapter 2 discusses the properties and characteristics of uranium, including its radioactivity and
chemical behaviour. It also outlines the life cycle of u ranium from the mining of uranium ore to the
production of enriched and depleted uranium hexafluoride (UF6). This chapter provides useful
background information, in particular section 2.2.3 describes the characteristics of DU used by the
DoD (US Department of Defence). It notes that DU may have trace amounts of 236U, which is not a
naturally occurring uranium isotope, but is sometimes present as a by-product of nuclear fission in
uranium derived from nuclear fuel (recycled) uranium. The mass fraction given in table 2.4
(0.0003%) is used here (appendix 1, table 2).

E2.3 The DU life cycle in the army

Chapter 3 considers applications (especially military) of DU, the regulations, policies and procedures
that the Army follows relating to DU weapon systems, and the Army’s radiation protection
program for DU. Non military uses include:
• neutron detectors;
• radiation detection and shielding for medicine and industry;
• shielding in shipping containers for radiopharmaceuticals, other radioisotopes, and spent
nuclear reactor fuel rods;
• components of aircraft ailerons, elevators, landing gear, and rotor blades;
• damping weights etc to suppress vibration during petroleum exploration;
• counterbalance weights in radar antennae and ballast in satellites, missiles, and other craft.

The US Army used DU alloy in three KE penetrator cartridges, shown in figure 3-3:
• 25-mm cartridge, weight 1.0 lbs (0.45 kg), DU penetrator weight 0.2 lbs (0.09 kg); to be used
in the Bradley fighting vehicle;
• 105-mm cartridge, weight 40.8 lbs (18.5 kg), DU penetrator weight 8.5 lbs (3.9 kg); used by
M1 and M60 tanks;
• 120-mm cartridge, (cross-section shown in fig. 3-1) weight 48 lbs (22 kg), DU penetrator
weight 10.7 lbs (4.9 kg); used by M1A1 and M1A2 Abrams tanks.

DU is also used in armour on the M1 series heavy armour (HA) tanks. Small amounts of DU are
used as epoxy catalyst in two types of mine. By 1994 more than 1.6 million DU penetrators had

3
been produced for tank ammunition and 55 million for small calibre (20-, 25- and 30-mm)
applications. More than 99% of the small calibre production was for the US Air Force (30-mm
GAU-8).

The chapter includes a description of sites licensed to possess DU, including the firing ranges. In
section 3.6.1 it notes that in 1979 the US NRC (Nuclear Regulatory Commission) prohibited
destructive testing that released airborne radioactive material to unrestricted areas. Hard target
testing was subsequently conducted in enclosures. Section 3.6.5 gives an indication of how long
DU has been in use; the Army Research Laboratory ARL-Aberdeen has conducted DU experiments
since the Army’s early use in the 1950s.

E2.4 Combat and post -combat DU issues

Chapter 4 describes and discusses:
• the Army’s use of DU-containing weapons during Operation Desert Shield/Desert Storm;
• the effects of a DU-penetrator round striking a vehicle, based on information obtained from
friendly fire incidents;
• the number of rounds used in combat and practice;
• how damaged and destroyed vehicles were handled, and the problems that personnel
encountered in dealing with DU contamination.

The key points are as follows.

DU-penetrators have a ‘sharpening effect’ upon impact that allows greater penetration through
armour (Hartline 1993, Danesi 1990). Weapon testing shows that when a DU round penetrates an
armoured vehicle, it may pass completely through the vehicle or ricochet around and fragment
inside the vehicle.
“When a kinetic energy round penetrates a vehicle, it contaminates the vehicle interior
with dust and fragments. Metal fragments from the penetrator and the vehicle’s hull can
scatter inside the vehicle, killing and injuring personnel, destroying equipment, and causing
secondary explosions and fires (figure 4-2). As much as 70 percent of a DU penetrator can
be aerosolized when it strikes a tank (Fliszar et al 1989). Aerosols containing DU oxides
may contaminate the area downwind. DU fragments may also contaminate the soil around
the struck vehicle (Fliszar et al 1989).” (We were unable to obtain the report by Fliszar et al
1989, but a summary is given in OSAGWI (2000) (see section E4.19), and further details in
CHPPM (2000).)

Also relevant to assessments, it is noted that most DU penetrators fired in combat will be in one of
three places:
• in or near target vehicles: 80 – 90% of tank rounds fired will hit the target and remain in or
near it;
• on the soil surface: projectiles that miss the target often ricochet, skipping across the ground;
they usually land within a mile or two of the target;
• buried in the soil: some projectiles will penetrate the ground; the percentage of buried rounds
depends on engagement angles and ranges, soil types and terrain.

E2.5 Previous DU studies

The introduction to chapter 5 notes that health and environmental risks of DU arise from
internalisation of metal fragments and oxides generated by impacts, fires and oxidation of spent
penetrators in the environment. Quantification of the risks requires knowledge of the quantities

4
and properties of the oxides and how these change with time.

The report notes that: “The Army has conducted many tests to determine the characteristics of
particles produced by hard and soft target impacts and by fires involving DU munitions and armor.
Unfortunately, data that the AEPI reviewed did not contain the attributes required to estimate
inhalation potential or environmental transport. AEPI is conducting an exhaustive review of existing
particle data to better define data gaps.” (The results of such a review would have been valuable to
our assessment, especially as AEPI would have greater access to information than we had.
However, we have not located it, or seen any further reference to it. The title of a report by
Parkhurst et al (1995) “Evaluation of DU Aerosol Data: Its Adequacy for Inhalation Modeling” is so
closely related that perhaps it is the review in question. We were unable to obtain a copy of the
report by Parkhurst et al but a summary is given in OSAGWI (2000) (see sections E4.19 and E4.20).

Sections 5.1 and 5.2 note that many tests have been carried out by or for the US Army that provide
information relevant to assessing potential health hazards from DU use. “Three investigation topics
are particularly significant:
• radiological doses received by crews in tanks loaded with DU munitions (ARDEC 1990,
Parkhurst et al 1991);
• DU aerosols generated by DU munitions (40 CFR 61, Jette et al 1990);
• DU aerosols resulting from the impact of various types of munitions on DU-armored tanks
(Fliszar et al 1989).”

It does, however, make two general criticisms:

• a lack of intercomparison between results of different tests or development of a predictive
model. “For example, AEPI researchers identified four studies of DU aerosolization during
munitions and armor testing that used similar experimental methods (Fliszar et al 1989,
GenCorp Aerojet 1993, Gray 1978, Haggard et al 1986). Although the studies were otherwise
excellent, none of the more recent studies compared their results with those from earlier
studies. This comparison is necessary to check the validity of results and to force researchers to
critically review differences and develop explanations for them. This step is essential in
developing a fundamental understanding of particle characteristics and in developing predictive
models.”
• a lack of peer-review of studies.

E2.6 Health issues associated with US Army use of DU

Chapter 6 discusses the state of knowledge concerning DU’s chemical and radiological toxicity,
medical protocols for DU-wound management, and DU-awareness training. It includes
consideration of the potential exposure of workers and soldiers, and gives information on external
dose-rates relevant to storage facilities and vehicles. It notes some less obvious minor pathways
such as from the gun tube into the crew compartment when a round is fired.

E2.7 Environmental risks associated with US Army use of DU and ways to reduce their
long-term effects
Chapter 7 considers the transport and fate of uranium released to the environment, and the
remediation of contaminated sites.

E2.8 Findings and conclusions

Chapter 8 makes a number of recommendations. The most relevant to our assessment are (section
8.2.3, p191):

5
• “Review DU-particle data from Army studies and elsewhere to determine data gaps.
• Develop and conduct experiments to generate the requisite data to fill these gaps. Data on DU-
chemical species, mass-mean size, surface-mean size, size distribution, specific gravity by
species and particle shape are required to support transport and risk models.
• Develop a better understanding of DU particles generated in fires or from hard-target or soft-
target impacts.”

E3 Rao and Bhat, 1997

Depleted uranium penetrators – hazards and safety. Defense Science Journal 47, 97–105.

This is a relatively concise review of the subject. It describes the material, the use of DU in KE
penetrators, including a list of 10 types of ammunition using DU penetrators from the USA, UK and
France, and compares it with tungsten. It covers uranium hazards: chemical and radiological,
medical and safety aspects. The abstract notes that “Open air firing can result in environmental
contamination and associated hazards due to airborne particles containing essentially U3O8 and
UO2”. Section 4.2 ‘Inhalation’ refers specifically to studies on test firings of 105-mm ammunition
with penetrators containing about 3.4 kg DU (Patrick and Cornette 1978, Glissmeyer and Mishima
1979, see annexe G). Based on those studies, it notes that when they impact against armour plate
targets, a cloud of airborne particles is generated by the spontaneous ignition of the fragments as a
result of a combination of shock and frictional heating. The particles range from microfragments of
diameters greater than 50 µm to submicron particulate aerosols. It reports estimates by Glissmeyer
and Mishima (1979) that approximately 2.4 kg of airborne DU was generated by each test firing,
out of which about 75% was U3O8 and the rest UO 2. It was observed that about half of the
airborne DU immediately above the targets was respirable, U3O8 being predominant in the
respirable range. About 43% of the respirable DU dissolved in simulated lung fluids within seven
days. Reference was also made to Hansen et al 1974.

E4 OSAGWI, 2000

Office of the Special Assistant to the Deputy Secretary of Defense for Gulf War Illnesses. Exposure
Investigation Report, Depleted Uranium in the Gulf (II), December 2000 at www.gulflink.osd.mil in
the Environmental Exposure Reports Section. A first interim report (OSAGWI 1998) was issued in
August 1998.

This is a very comprehensive review (nearly 300 pages) of DU in the context of the Gulf War. The
following sub-sections use the same headings as the subdivisions of the report, to show its scope,
but this summary is selective in selecting items most relevant to our assessment. A series of
appendices are labelled Tab A – S.

E4.1 Overview (I)

Its purposes are (page 6) “to determine whether DU posed an unacceptable health risk to American
forces and whether personnel had been adequately trained to deal with this risk. To accomplish
these objectives, the report examines the documented incidents of DU exposure and discusses what
is currently known about the potential health effects from them.” Possible exposures are classified
into three Levels (I, II and III), encompassing 13 activities. (These seemed to provide a reasonable
basis for our assessment. However, whereas the OSAGWI report is concerned specifically with the
Gulf War, ours is wider, including possible future conflicts. Table E1 below is based on OSAGWI
2000 table 1 (Incident summary), but references to specific US personnel and Gulf War situations

6
have been removed.)

Level I includes the highest exposures, “immediate and direct”. Level II includes direct contact with
DU dust, but after combat. Level III includes generally short-term, very low exposures. The three
levels are discussed further in chapter IV, and in detail in Tab G. (OSAGWI considers the
USACHPPMM’s Level II and III assessments adequate, but not those for Level I.) “The amount of DU
present, route of entry, solubility, particle size, other physical and chemical factors, and toxicity
determine potential health effects. The US Army Center for Health Promotion and Preventive
Medicine (USACHPPM) completed its health risk characterisation of DU in the Gulf War after we
published our initial environmental exposure report on DU. They reassessed earlier Level I estimates
the General Accounting Office (GAO) called into question, [2] and developed Level II and III
estimates. Although more refined than their original estimates, USACHPPM’s new Level I estimates
rely on the same test data as used previously. USACHPPM employed statistical tools to develop
upper and lower limits for these Level I exposure scenarios. To improve the reliability of these Level I
estimates, OSAGWI has directed and funded the US Army to further evaluate DU aerosol
concentrations inside combat vehicles penetrated by DU rounds. [3] In the meantime, the Baltimore
Veterans Affairs (VA) Medical Center’s comprehensive medical follow-up program provides the
most important health assessment for Level I exposures. The VA’s studies of these Level I veterans
have shown no untoward medical effects to date from depleted uranium’s radiological or chemical
toxicity. USACHPPM’s risk assessments for the Level II and III scenarios are based on much better
Department of Defense (DoD) experimental data and indicate that the radiological and chemical
risks for these events are well within current regulatory limits for industrial workers. These results
for participants in all levels confirm our initial scenario classification.”

Table D1. Battlefield exposure levels

Level Examples
I Soldiers who:
were in or on a vehicle when a DU munition penetrated it;
entered vehicles to rescue occupants immediately after DU impacts.
II Personnel who:
removed equipment and munitions from vehicles struck by DU munitions;
performed maintenance on or recovered items from vehicles struck by DU
munitions;
inspected vehicles struck by DU munitions to determine reparability;
examined combat vehicles damaged and destroyed by DU munitions;
processed damaged equipment, including some struck by DU munitions;
were radiation control team members;
were exposed to DU during cleanup operations after a fire at a vehicle and
munitions storage area.
III Personnel who were exposed to smoke from:
burning DU rounds in a fire at a vehicle and munitions storage area;
burning vehicles with DU armour and/or munitions on board;
DU-struck equipment.
Personnel who entered DU-contaminated equipment.

E4.2 Methodology (II)

The study used a risk assessment methodology based on that of the USEPA (Environmental
Protection Agency), which involves four steps:

7
• hazard identification—who was exposed, and how?
• toxicity assessment—what are the known medical effects of human exposure to DU? At what
levels of exposure do these effects occur? How can the effects be diminished?
• dose assessment—how much DU were soldiers exposed to? How much did they take into their
bodies? What chemical or radiological doses do these intakes represent?
• risk characterization—using validated toxicity and dose information, what medical effects can
be anticipated? How serious are they? What is the risk they will occur? How can the effects be
communicated to those affected?

This assessment involved several organizations:

• the Office of the Special Assistant for Gulf War Illnesses (OSAGWI) for hazard identification and
risk characterization;
• the RAND Corporation for toxicity assessment (section E5);
• the US Army Center for Health Promotion and Preventive Medicine (USACHPPM) for dose
assessment (exposure and risk assessment). Since the chemical intakes and radiological doses
were not directly measured when the incidents occurred, USACHPPM used the best available
data, combined with scientific and engineering principles and data from relevant tests. Specific
activities included: reviewing test data on DU’s behaviour during fires and impacts with
armour; evaluating the usefulness of these data in modelling the amount of DU a soldier might
take in and retain in the body through inhalation or ingestion; identifying data gaps; and
estimating the radiological and chemical doses for each of the 13 activities involving possible
DU exposure. (These activities correspond closely to our assessment.)
• the Department of Veterans Affairs (VA) and Department of Defense DU medical surveillance
programs for medical follow-up.

E4.3 Depleted uranium – a short course (III)

This section discusses DU’s chemical, physical and radiological properties, the ways those properties
may affect human health, and the principles and standards for protecting soldiers and the public
from harm.

Section III.A.2 notes that: “In the Gulf War environment, DU entered the body through inhalation,
ingestion, or wounds —in the form of uranium metal (from flying fragments and unoxidized DU)
and uranium oxides (mostly depleted triuranium octaoxide (U3O8) but also depleted uranium dioxide
(UO2) and depleted uranium trioxide (UO3) from DU impacts on target vehicles or fires”. (Source
documents we have seen refer to formation of U3O8 and UO2 in varying proportions in impacts and
fires, but not UO3, which dissolves much faster in the lungs (annexe A, table A5). The statement is
referred to Harley et al (1999 section E5). Harley et al (1999) in turn (section 2 page 12) refers this
to CHPPM (1998), which we were unable to obtain. CHPPM (2000, P J-90) states that “Only DUO2
and DU 3O8 have been observed from either hard-target tests or fires. However, because DUO3 has
been found in some of the range studies and is considered to be the most soluble of the three DU
oxides, it was chosen to model the soluble fraction.”)

Section III.B.1 notes that: “The Department of Energy (DOE) recently reported that the DU stock it
provided to DoD for manufacturing armor plates and munitions may contain trace levels (a few
parts per billion) of transuranics (neptunium, plutonium, and americium). Transuranics are
radioactive elements with higher atomic numbers (more protons and electrons) than uranium. To
verify the level of transuranics in the DU stock material received from DOE, the Army tested
representative samples from various batches of DU stock used to manufacture DU armor plate.
From a radiological perspective, the transuranic contamination in DU armor contributed an

8
additional 0.8 percent to the radiation dose from the DU itself. Scientists consider this insignificant
considering the very low radiological hazard associated with the primary material, DU. [51] While
this issue has received considerable attention at certain DOE facilities and in the press, the
implications for DoD are minimal—the quantities are so small they add very little to the radiation
dose from depleted uranium itself. Both DOE and DoD concluded that measures designed to
protect personnel from the DU itself are more than adequate to protect them from the trace
quantities of transuranics.”

Tab C includes a similar statement: “The Department of Energy has recently reported that the DU
used by DoD in its armor plates (found only on “Heavy Armor” Abrams tank models) may contain
trace levels of transuranics (neptunium, plutonium, and americium) and fission products
(technetium-99). The DU used in munitions may also contain these materials. The military services
are testing the stocks of DU munitions and parts. The levels of transuranics and fission products
found during testing of the material used for producing armor packages are in minute quantities
(the picocurie/gram range) and result in less than a one percent increase in the internal radiation
dose. These evaluations indicate that measures designed to protect personnel from the DU itself are
adequate to protect them from the traces of transuranics and fission products as well.”

(The reference [51] given is “Letter to US Nuclear Regulatory Commission from US Army Tank-
Automotive and Armaments Command, Subject: "Nuclear Regulatory Commission (NRC) Licenses
SUB-1536, SUB-1564, SMB-141 and SM-0179," January 20, 2000, Enclosure 1, p. 13-14.” A report
of such measurements was recently obtained (Bhat 2000). On the basis of the measurements it
reported, it was confirmed that following an intake of DU the transuranics and 99Tc contributed
less than 0.1% of the dose from the uranium isotopes (annexe D, section D2).)

E4.4 Potential health effects from DU use in the Gulf Theater, 1990-1991 (IV)
This chapter provides a summary of the dose and risk assessment methods used by USACHPPM.
Section IVA outlines the methodology. Tab O discusses the intake and dose assessment process in
more detail. It notes that: “Chemical intakes and radiation doses from DU entering the body
cannot be easily measured directly. Bioassay methods, such as measuring uranium in urine or
inhaled air, reasonably estimate doses. Except for several service members who unloaded munitions
from a tank after a fire had ‘cooked off’ its rounds, DoD did not promptly measure DU intakes or
doses. So initial exposure estimates and risk assessments relied on the next best method—
estimating the dose and risk derived from scientific principles and data measured under conditions
similar to those in the Gulf War.” (We have necessarily taken a similar approach.) Further
information about the group that was monitored is given later in the chapter; see annexe C,
section C3.6.

Sections IVB–D consider Level I, II and III exposures. More details are given in Tabs G, H and I. Tab I
describes the post -war fire and cleanup at the Camp Doha motor pool which on July 11, 1991,
destroyed three M1A1 tanks loaded with DU rounds and several hundred DU rounds stored nearby.

Initial assessments for Level I, based on sparse test data for DU behaviour inside penetrated
vehicles, indicate soldiers inside vehicles when they were struck by DU munitions and those
rescuing fellow soldiers in stricken vehicles could have received exposures higher than some
guidelines in a limited number of cases involving very conservative scenarios. Upper limit estimates
of intake by inhalation and ingestion were about 30–240 and 40–70 mg, respectively. Exposures
would be lower for lightly armoured vehicles ( eg Bradley Fighting Vehicles) than for tanks, because
the penetrator tends to pass through without fragmenting. They would also be lower for rescuers if

9
time had passed allowing the initial high concentration to fall or disperse. According to OSAGWI,
USACHPPM used the best a ir concentration data available, contained in Fliszar et al (1989). (We
were unable to obtain this report. It is summarised in Tab L, Report Summary #27 (see below).)
However, it is also noted that the General Accounting Office (GAO) questioned the reliability of this
data (section E4.24), and the U.S. Army was directed to conduct a programme of further tests,
which is expected to be completed in Fiscal Year 2002.

For Level II, USACHPPM evaluated the exposures in two groups, Field Units and Camp Doha. Tab O
provides details. For Field Units, it focused efforts on inhalation and ingestion during re-entry of
DU-contaminated vehicles and activities around and near contaminated vehicles. Because
contamination levels and exposure times varied widely and were known only in general, it
developed chemical and radiation dose factors based on one hour of exposure in a vehicle.
Multiplying those factors by the exposure time in each vehicle and the number of vehicles entered
could produce estimates for Level II intakes and doses. Estimates of intake by inhalation and
ingestion were 0.025 and 0.057 mg per hour, respectively. Considering the number of
contaminated US vehicles, USACHPPM estimated upper limits of DU intake by inhalation and
ingestion at 2.3 and 5.3 mg DU respectively.

For Camp Doha, USACHPPM requested Pacific Northwest National Laboratory (PNNL) to review and
analyse data from burn tests of DU munitions to assess the possible exposures to Camp Doha
personnel, because PNNL had conducted many DU tests during the past two decades. PNNL
determined the pathways, (including inhaling DU oxides from smoke generated by the fire or
resuspended from contaminated surfaces, and ingesting DU oxides from contaminated surfaces),
and assessed the intakes of DU. PNNL calculated for each category (OSAGWI 2000, table 6) DU
concentrations in the kidneys (all below 0.1 µg per gram kidney) and radiation doses (all below 0.1
rem, ie 1 mSv, committed effective dose equivalent).

Level III represents individuals who received relatively fleeting DU exposures from climbing on or
entering combat vehicles to assess damage, remove equipment or collect souvenirs, and persons
exposed to the smoke from fires involving DU, notably Camp Doha. Besides the Level II personnel
involved in cleaning u p, hundreds more may have received short-term exposure to the smoke from
burning DU munitions. Many Level III exposures involved the same scenarios and routes of entry as
those in Level II but shorter exposure times. To simplify the assessments, USACHPPM classified Level
III participants into Field Units and Camp Doha personnel. For Field Units, estimates were made for
intake by inhalation and ingestion for one hour of exposure to one vehicle, summarised in table E2.
This is based on table 7 of OSAGWI 2000, which also gives estimates of resulting kidney
concentrations and doses.

Table D2. Summary of Level III Field Unit Intakes (mg per vehicle per hour)
Scenario Inhalation Ingestion
Smoke from a burning tank 0.0028 –
Entry of a burned tank 0.025 0.057
Entry of DU-damaged or destroyed Iraqi vehicle 0.0057 0.057
Downwind of DU penetrated vehicle 0.0044 –

USACHPPM developed a hypothetical, composite scenario that involves a cavalry scout who
contacted DU in the field, and which produces a total DU intake of 0.065 mg:
• taking 20 minutes to pass one burning Abrams Heavy Armor tank;

10
• entering 7 enemy vehicles hit by DU and spending 7 minutes in each vehicle without
contaminating his hands;
• entering one enemy vehicle for one hour, contaminating his uncovered hands, which he then
did not wash for two days; and
• taking 10 minutes to pass at a distance of 80 meters through the smoke of 3 burning enemy
tanks struck by DU.
For Camp Doha, PNNL estimated that the highest exposures any person could have experienced
occurred one kilometre downwind of an elevated release and produced a maximum kidney
concentration of 2.8 x 10 -7 µg DU/g tissue and a maximum radiation dose (CEDE) of 3.0 x 10 -6 rem
(3.0 x 10-5 mSv).

E4.5 Follow-up (V)

This Section discusses environmental assessments of DU battlefield contamination, recent
environmental studies of various DU munitions, the results of current medical studies, and ongoing
and planned research.

Section A (Environmental Assessments) provides information relevant to the amount of DU likely to

be dispersed. When a penetrator fired from a tank or aircraft impacts, only a fraction is aerosolized.
The Army fired 9,552 DU tank rounds (approximately 50.55 tons) while A-10s fired 783,514 30-
mm DU rounds (approximately 259 tons). The tank rounds were much more likely to hit their
intended target than the 30-mm rounds, though the exact number of 30-mm rounds that struck
targets in the Gulf War is unknown. Simulations conducted before the Gulf War indicate only a
small percentage of the A-10 aircraft rounds (less than 10%) actually hit the target, eg during 9
passes with the A-10s firing 2-second bursts, only 93 of 957 rounds actually hit the targets. The DU
munitions fired from aircraft can miss the target entirely, hit the target and ricochet, or embed in
the target without penetrating. Each of these circumstances leaves the penetrator almost entirely
intact and produces little or no aerosol or fine particles.

Section B (Post-Gulf War Developmental Testing and Evaluation of DU Munitions) reports that the
only new DU munition the US has fielded since the Gulf War is the M919 25-mm APFSDS-T
cartridge, which entered service in 1995 for use in Bradley Fighting Vehicles. Test results were
consistent with those from hazard-testing other DU munitions. The testing report concluded:
“There was no indication that any measurable DU became airborne as a result of the External Fire
Stack Test”. During hard impact testing, less than 10% of the DU penetrator was aerosolised. Less
than 0.1% of the initial mass of the penetrator was in the respirable range. Of the oxide formed,
83% was insoluble. The reference for this was Parkhurst et al (1990) (Tab L Report #29). (We did
not obtain a copy of this report.) In 1994 the Army conducted a burn test on a Bradley Fighting
Vehicle loaded with 1125 DU rounds. This was the first burn test to involve a real vehicle fire, rather
than a stack of wood etc (see annexe H). “Although the fire and subsequent explosions released a
small amount of DU oxide, the testers detected only trace amounts on the air monitoring filters
placed at various distances from the Bradley.” The reference for this was Parkhurst et al (1999) (Tab
L Report #42). (We did not obtain a copy of this report.)

It reports that the Army also conducted DU hard-impact aerosolisation tests on various foreign
armoured vehicles in 1995. “Several technical and procedural difficulties, eg sampler failure,
seriously affected the data and limited the conclusions that could be drawn from these tests.
Nevertheless, the Army found:
• a DU penetrator’s impact against an armored vehicle generates DU aerosols containing
particles of respirable sizes inside the vehicle. The concentration of DU aerosol decreases with

11
 time, but measurable concentrations of respirable particles remain suspended inside the vehicle
hours later;
• measurable quantities of DU oxide particles can resuspend when individuals re-enter the
vehicle; the resuspended aerosols contain particles of respirable sizes.”
The reference for this was ARDEC (1998) (no summary in Tab L). (We did not obtain a copy of this
report, which looks particularly useful in relation to assessing intakes following entry to
contaminated vehicles.)

Section C (DoD and VA Medical Surveillance Programs for Gulf War Veterans; a more deta iled
description is in Tab P) describes the studies on veterans with embedded DU fragments (McDiarmid
et al 2000, Hooper et al 1999).

It also describes urine monitoring on veterans (see also annexe C, sections C2.3 and C3.6). Some
whole body counting and urine testing was carried out in 1992 and 1993. Results are described as
being within normal ranges. In 1998, the medical follow-up programme expanded to evaluate all
veterans who were in or on vehicles struck by friendly fire and those who worked around DU-struck
vehicles or burned vehicles containing DU (Level I and Level II). The purpose was to measure the
uranium in veterans’ urine, to verify that they had normal amounts of uranium in their bodies. Level
III veterans concerned about their possible DU exposure can also obtain a DU medical evaluation.
Participants in this expanded program submitted a urine sample, which was evaluated for uranium.
For amounts higher than 50 nanograms of uranium per gram of creatinine, the facility retested the
veteran. As of September 2000, of the 309 veterans tested, 11 initially tested higher than the
screening guidelines. Seven veterans have had follow-up samples tested, three of which were
elevated a second time.

Section D (Medical Testing by Other Laboratories) notes that media reports have cited claims of
elevated uranium in urine samples from veterans in the United States, the United Kingdom and
Canada based on data from US and Canadian researchers. Recognising that the discrepancies
between the government’s and outside laboratories’ test results concern veterans, OSAGWI
initiated a laboratory assessment study of approximately 10 laboratories that have analysed Gulf
War participants’ urine.

Section E (Postwar Research and Literature Reviews) outlines research related to embedded DU
fragments, and three major scientific reviews of the toxicology of uranium and depleted uranium
that were published since the initial OSAGWI report was published in 1998: ATSDR (1999), Harley
et al (1999) and Fulco et al (2000).

E4.6 Lessons learned and recommendations (VI)

E4.7 Conclusion (VII)

E4.8 TAB A – Acronyms, Abbreviations and Glossary

E4.9 TAB B – Units Involved

12
E4.10 TAB C – Properties and Characteristics of DU

E4.11 TAB D – Methodology

E4.12 TAB E – Development of DU Munitions

E4.13 TAB F – DU Use in the Gulf War

E4.14 TAB G – DU Exposures in the Gulf War

This gives details on Level I, II and III exposures. It also includes a few notes that DU was not the
only potential radiological or toxic hazard associated with damaged Iraqi vehicles. For example:
• in section B6: “The AMCCOM personnel also surveyed captured Iraqi equipment being
prepared for shipment to the US. According to the person in charge of the survey operation,
the most acute radiological hazard on these Soviet-built tanks was the radium used in their
gauges, which were often leaking…”;
• in section D3: “War trophy hunting became a problem. Many soldiers and leaders did not
recognize the hazards in war trophy hunting. Booby traps, radiation contamination from
depleted uranium, and unexploded ordnance combined to make this practice dangerous. In
addition, units wanted to take home pieces of enemy equipment. This equipment can have
gauges and other items that contain radium-226. We also found some Iraqi tanks wi th
asbestos blankets…”.

E4.15 TAB H – Friendly -fire Incidents

E4.16 TAB I – The Camp Doha Explosion and Fires (July 1991)

E4.17 TAB J –Tank Fires

This describes four accidental tank fires. The fourth (section D April 13, 1991) is significant in that
the crew had urine samples collected at the time, the only reported urine samples measured for
uranium directly after exposure. (The next were in 1992, see Tab P.) This is discussed in annexe C,
section C3.6.

E4.18 TAB K – DU Notification and Medical Follow-up Program

E4.19 TAB L – Research Report Summaries

“This tab lists chronologically some of the major research on military use of depleted uranium (DU).
Though not all -inclusive, this list indicates the depth and breadth of research conducted to date.”
This provides a useful list of relevant reports, but unfortunately many are not in the public domain,
and the report (here and elsewhere) does not state which are, and which are not. (We have
extracted information below from a few of these which summarise particularly relevant reports that
we were unable to obtain.)

Report Summary 10: Gilchrist et al (1979) PNL-2944

(This was obtained, and is referred to here (and elsewhere) as Glissmeyer and Mishima (1979). See
Annexe G.)

Report Summary 11 Gilchrist et al (1979)

Battelle (now PNNL) released this report to the US Army in 1979, but it was not then published. It

13
was published in June 1999 because of current interest in DU and the uncertainty associated with
the frequently cited 70% aerosolisation figure. That value comes from Glissmeyer and Mishima
(1979). ((In OSAGWI 2000, this is cited as Gilchrist et al 1979 and outlined in Research Summary
10.) This report was not cited in OSAGWI 1998, and we first saw reference to it in OSAGWI 2000.
A copy has been requested (November 2000) but has not yet arrived (May 2001).)

This report is a follow-up to the hard target impact testing of the M774 (Glissmeyer and Mishima
1979) with improved test conditions and sample collection. In this test, 105-mm M735E1 cartridges
were fired at a series of three armour plate targets located about 200 m away. The impact and
penetration of the projectiles caused a shower of fragments and considerable airborne particles.
The overall results included the following:
• The airborne fraction of the penetrator from target impact was in the range 17–28%. The
uranium air concentrations near the target were above the then maximum permissible
concentration in air (MPC a) for occupational exposure for the first 5 minutes post-shot. (The
MPCa was calculated to be 0.28 mg m–3). The airborne level of DU remained elevated for at
least 15 minutes when the surface was dry, but fell below the MPC within 15 minutes when
the surface was wet.
• At 4 to 5 m, the uranium concentrations were below the MPCa within 10 minutes and
remained so irrespective of surface wetness or of work activity near the target after the 15
minutes post-test mark.
• The wetness of the ground surface had no clear impact on the airborne cloud volume or the
total mass of material made airborne immediately after a test firing.

There were two separate particle size measurements at the target accessway. The authors believed
the MMAD of 2.1 µm from 4-stage high-volume cascade impactors (operated at 20 cubic feet per
minute, cfm) was probably better than the 5.8 µm MMAD value calculated using the low-volume
eight-stage cascade impactors flow rates (operated at 1 cfm (4.9 x 10-4 m3/sec = 29 lpm)).

Report Summary 21 Haggard et al (1986)

This followed up an earlier test (with a wooden shipping container) to evaluate a new metal
shipping container. Relevant test findings were:
• about 9.5% percent of the total DU in the cores was converted to oxide during the fire;
• the oxide was predominantly U 3O8;
• the fraction of generated oxide small enough to be suspended in air and carried by the wind
was 0.2 – 0.6%;
• the fraction of generated oxide small enough to be inhaled was about 0.07%;
• in simulated lung fluid 4% dissolved within 10 days and the rest was essentially insoluble;
• during the test, winds were relatively calm. No DU was detected by downwind air samplers
(detection limit of 1 mg DU).

Report Summary 27 Fliszar et al (1989)

This test evaluated DU aerosol levels generated inside and outside a heavy (ie DU) armour Abrams
tank hit by various types of rounds. The test also evaluated particle size distributions of DU puffs
near the point of impact and within 100 m from the tank, resuspension levels within 100 m of the
tank, and DU contamination in air from a DU munitions-loaded burning M1A1 heavy armour tank
after being struck.

The seven tests used the following rounds: 120-mm APFSDS KE (kinetic energy)-Tungsten; 120-mm
Heat-MP; 100-mm AP-C steel rod; anti-tank mine; 120-mm APFSDS KE DU (test 5A); 120-mm
APFSDS KE-Tungsten (test 5B); and ATGM equivalent (test 6B).

14
In evaluating the test data, it is important to distinguish between the aerosols typically generated as
puffs from impact and aerosols generated from a fire plume involving DU penetrators. In all tests
the highest fallout levels occurred on the test pad within 5 to 7 metres of the target, but after
several tests heavy armour material was blown out more than 76 metres.

Based on the test data, exposures from passing clouds generated at impact were considered
insignificant beyond 100 m. The maximum estimated intake of DU from a passing airborne cloud at
200 m from the target vehicle was about 1 microgram ( µg). Within 100 m, but outside the cloud
path, including air samplers within 5 to 10 m of the target, results were also insignificant. Air
sample results in the cloud path varied; the air samplers recorded the highest amount, 280 µg (an
acute exposure) 10 m from the target, with little additional intake after the puff passed. Air
sampling results for test #7, which caused a fire that consumed the vehicle, still were within the
intake limit even though the air samplers also were exposed to the smoke plume of the fire.
Cascade impactor data for smoke puffs generated at impact showed that the particles in the cloud
were primarily (about 80%) respirable.

The resuspension air sampler results 10 to 100 m from the target revealed, at least for this test,
resuspension was not a problem. The highest amount recorded was “1.7×10-14 microcuries/ml”
(6×10-4 Bq/m3 = 0.05µg/m3), well within the limit for airborne uranium. The resuspension samplers
were started after re-entry to the test area to test for resuspension. In this case, resuspension
includes both the material on the ground and loose material from the target itself that may be
resuspended.

An initial re-entry team member wore a personal sampler in the breathing zone to evaluate
resuspension at the test pad and while climbing inside the crew compartment. All resuspension
results were within acceptable limits except for the Hellfire equivalent test, in which the team
member re-entered after the fire and collected samples primarily from inside the crew
compartment. The report indicated a penetrator might have been ejected from one of the storage
compartments into the crew compartment and then completely oxidized during the test. Even so,
the report stated the airborne concentration was slightly higher than the soluble 238U limit, and the
insoluble 238U limit (5×10-12 microcuries/ml = 0.2 Bq/m3 = 15µg/m3) probably was appropriate. Based
on the insoluble 238U criteria, all resuspension data would be within acceptable limits.

Researchers also sampled interior air during the three last impact tests, when breakthrough into the
crew compartment occurred (see annexe C, section C2.2).

Report Summary 29 Parkhurst et al (1990)

The 25-mm, M919 cartridge was not used in the Gulf War, but a summary of the testing was
included for comparison with tests on the cartridges that were used. Tests included the Stack Test,
which evaluates detonation propagation, and the External Fire Stack Test, which evaluates the
cartridges’ explosiveness and fragmentation resulting from setting fire to boxes of cartridges. The
M919 was also tested against hard armour and wood/masonry targets to determine the extent and
nature of DU aerosols created. Relevant results included the following.
• In the External Fire Stack Test ~35% of the total DU used was oxidized. Between 0.1% and
0.2% of the oxide was within the respirable range. The lung solubility analysis of the DU oxide
determined that 92.6% was insoluble and 6.8% was slightly soluble. There was no indication
that any measurable DU became airborne.

15
• In the hard target impact testing <10% of DU was made airborne. Less than 0.1% of the initial
DU penetrator weight was within the respirable size range. About 17% of the oxide present in
the smallest size fraction was soluble while the remaining 83% was insoluble.

Report Summary 31 Jette et al (1989)

This study characterised particulate levels with both complete and partial penetration of the armour
after hard impact. Two DU rounds were tested. Two non-DU rounds were also tested to evaluate
DU resuspension during hard impact tests. The results were questioned because the amounts
aerosolised were estimated to be only 0.2–0.5% for the M829A1 and 0.02–0.04 percent for the
XM900E1, two orders of magnitude lower than expected from previous trials. This study stated
that it was highly unlikely that more than 10% of the DU by weight aerosolised on impact. Results
indicated that a high percentage of the respirable dust was soluble in the lungs: 24–43% was class
D material (dissolution half-time <10 days), and the rest Class Y (dissolution half -time >100 days).
The resuspension tests indicated most of the resuspended dust was non-respirable, consistent with
the theory that the enclosure’s filtering system removed most of the respirable dust. (From this, we
can only infer: (a) <10% of penetrator mass converted to airborne dust, consistent with Chambers
et al (1982); (b) Rapid dissolution fraction ~30%, consistent with Glissmeyer and Mishima (1979).)

Report Summary 32 Munson et al (1990)

This was a desk assessment to predict potential radiation hazards to personnel entering a site
where DU has hit a tank. The authors based it on a DU penetrator from a 105-mm round striking
and penetrating one side of an armoured vehicle (without DU armour or DU munitions on board).
They assumed that the impact would result in aerosolisation of 12–37% of the DU penetrator
mass, and that 44–70% of the aerosolised DU (mainly oxide) would be less than or equal to 3.3 µm
dae , and could be inhaled into the deep lung. The report estimated doses from internal and external
exposure and recommended precautions during clean up and recovery.

Report Summary 40 Parkhurst et al (1995)

“As its title implies, this study evaluated existing research data on the characteristics of DU aerosols
generated under various conditions, focusing on chemical composition, particle size, and solubility
in lung fluid. The report summarises more than 20 of Battelle’s own studies and 20 more studies
conducted by other researchers. Although the researchers cited several areas as needing further
research (eg resuspension and particle size distribution), the researchers deemed the data’s overall
quality adequate to conservatively estimate dispersion and health effects.” (This would have been
useful, both a source to locate the various studies, and for the opinions of those who carried out
many of them. More details are given in Tab M.)

Report Summary 42 Parkhurst et al (1999)

Although Bradley Fighting Vehicles (BFV) did not use DU ammunition during the Gulf War, this test
simulated an accident or attack that destroys a BFV loaded with 25-mm DU ammunition. Of the
1125 penetrators, 619 were recovered (full or partial): 80% within 2 m of the BFV, 98% within 50
m. It was estimated that 56% of total DU was recovered, and more was in the burned-out BFV and
surrounding soil. The total aerosolised could not be precisely estimated.

An extensive air monitoring network was employed consisting of downwind, high-volume air
samplers and fallout trays. The only air filters with measurable concentrations were at 30 m. Small
amounts were found on some fallout trays. The DU oxide recovered from inside the vehicle was
U3O8, with an AMAD ~13 µm, 33% <10 µm d ae and 6.5 % <3.3 µm d ae.

It was concluded that DU oxide was generated and some was small enough to be respirable, but
most of it was not available for transport under the test conditions.

16
E4.20 TAB M – Characterizing DU Aerosols
This appendix provides OSAGWI’s overview of the information on DU aerosol characteristics, based
on the various tests, many of which are summarised in Tab L.

It notes that various impact tests have generated widely varying amounts of aerosols. Glissmeyer
and Mishima (1979) reported on one of the first DU hard impact tests, and concluded that 70% of
the penetrator aerosolised. According to OSAGWI: “Although some have frequently cited this 70
percent figure, it is flawed and misleading—mainly because it was "back-calculated" from cloud
data and represented a worst-case scenario (ie an impact against a hard target, which was not
penetrated).”

Chambers et al’s (1982) report contradicted the 1979 test results: ~3% of the DU was aerosolized
(highly unlikely to be more than 10%). Approximately 44% of the aerosolised particles were less
than 1.1µm dae , and 70% were less than 7 µm, ie respirable particles.

Hard-impact testing by Jette et al (1990) of 120-mm and 105-mm cartridges produced different
results. This study characterised particulate levels after hard impacts with both complete and partial
penetration of the armour. Scientists questioned the results because only about 0.02–0.5% of the
DU aerosolised, approximately two orders of magnitude below expected values. Subsequently
researchers estimated that ~18% aerosolised, still much less than the 70% cited by Glissmeyer and
Mishima (1979). The respirable aerosol fraction (<10 µm dae) was 91–96% for the M829A1 and
61–89% for the XM900E1. For the respirable fraction 24–43% was class D material (dissolution
half-time of <10 days) and the rest was class Y material (dissolution half-time >100 days), agreeing
with other studies indicating a high percentage of the respirable dust from hard-impact testing was
soluble in the lungs.

Following recommendations in AEPI (1995), Parkhurst et al (1995) evaluated existing test data for
predicting aerosol exposures (summarised in Tab L Report #40, but more information is given here).
The report identified some technical problems in estimating exposure under various combat
scenarios. It briefly discusses DU aerosol generation scenarios:
• Fires. During a munitions ‘cookoff’, combustion is so rapid that little DU metal oxidises. In the
absence of explosions, few of the particles generated are small enough to be caught up in the
thermal currents. The oxides formed during a fire have very low solubility. Most particles
produced in a tank fire adhere to the interior walls, but openings can let particles escape into
the surrounding atmosphere.
• Vehicles punctured by projectiles. The amount of oxide formed during impact depends largely
on the target’s ‘hardness’. The heavier the armour, the more oxides will form as the DU
penetrator expends its kinetic energy piercing it. During the Gulf War, DU hits on lightly
armoured vehicles typically left golf ball-sized entrance and exit holes. DU aerosolisation was
limited unless the round struck the engine or a similar obstruction. Aerosolisation is enhanced
if the penetrator splits into fragments and the fragments remain inside the vehicle. Aerosol
levels inside the vehicle also depend on factors such as the number of open hatches and other
ruptures or openings. Eventually, particles from inside the tank either adhere to the tank’s
inside surfaces or release into the atmosphere through any opening. As they accumulate on
interior surfaces, the particles’ size and mass change.
• Entering contaminated vehicles. For emergency rescue personnel who enter a tank shortly after
impact, the aerosols generated at impact would be the primary threat. These impact aerosol
levels usually will be higher than those of resuspended DU particles remaining after the
aerosols in the tank have had time to settle or vent through open hatches etc. For battle

17
 damage assessment teams, recovery personnel, or souvenir hunters entering a damaged
vehicle, the primary threat is resuspended DU dust. Resuspension depends on the air
turbulence inside a vehicle and other conditions (eg oily surfaces reduce resuspension). Physical
activity inside the vehicle (eg lifting or moving equipment or personnel) would increase
resuspension.
• Inspecting and repairing contaminated vehicles. Entering contaminated vehicles to inspect and
repair them can cause significant DU resuspension. Some repair activities and cleaning
operations can also cause resuspension.
• Routine combat activities. DU penetrators that did not pierce the target or were deflected
could expose personnel. The penetrator would be hot enough to generate aerosols and oxides
would continue to form, even after it buried itself in soil. The report also cited potential
exposure to personnel near the target at impact or exposed to resuspended dust from
subsequent activities on, in, or near the target.

Two more recent tests conducted since the 1995 report raise questions about the nature and
extent of respirable particles generated during fires and hard-impact testing. In the first, DU
munitions were fired at Soviet armoured vehicles (ARDEC 1996). (This is also described in Section V.
Follow-up, B Post-Gulf War Development and Testing of DU Munitions, but the reference there is
ARDEC (1998).) The second was the burn test on a loaded Bradley Fighting Vehicle (Parkhurst et al
1999, Tab L Report Summary #42).

E4.21 TAB N – Gulf War Protective Guidance

E4.22 TAB O – DU Dose and Risk Estimates for the Gulf War Theater, 1990-1991
This appendix summarises OSAGWI and USACHPPM efforts to estimate and report intakes and
radiation doses from DU and to characterise their risks to health. This is summarised in chapter IV
4.4.4 Potential health effects from DU use in the Gulf Theater, 1990-1991 (section E4.4 above).
Much more detail is given in CHPPM (2000). This summary picks out only a few key points.

Wounding by DU fragments is not included, because soldiers in that category are participating in
the Baltimore Veterans Affairs (VA) DU Program. Inhalation was of most concern in each Gulf War
scenario that was considered. Doses from contaminated wounds were not assessed because the
NCRP (National Council on Radiation Protection and Measurements) is in the process of developing
a model.

CHPPM used current ICRP biokinetic models: ICRP publication 66 respiratory tract, ICRP publication
30 GI tract and ICRP publication 69 uranium systemic models (annexe A). The LUDEP (V2.06)
software developed by NRPB was used (Jarvis et al 1996), although it does not implement the ICRP
publication 69 uranium systemic model. (CHPPM analysis concluded that effective doses were
within 1% of those using all the current models.) Critical input parameters were breathing rate,
particle size and solubility class.

Input assumptions for dose calculation were:

• AMAD = 5 µm inhaled by a ‘mouth breather’ at 3 m3 per hour, simulating ‘heavy activity’;
• DU oxides from impacts are 17% type M and 83% type S as determined from the test data.
For transfer to kidney, CHPPM used ICRP publication 30 lung and systemic uranium models to
evaluate the fraction of intake that reaches the kidneys. For 5 µm AMAD, fractions are 6.42%
(class D), 1.74% (class W) and 0.34% (class Y). The analysis made the conservative assumption that
all the DU reaches kidney at the same time (this will lead to an overestimate, especially for class Y).

18
It assumed DU oxides are 17% class W and 83% class Y.

Consideration was given to uncertainties. The main ones relate to individual exposures, because of
lack of information on exposure times and conditions. Additional uncertainties relate to use of test
data obtained under conditions different from those in exposures, and uncertainties in the models
used.

For estimating Level I exposures, given lack of any air concentration measurements during the Gulf
War, CHPPM considered the best data source to be Fliszar et al (1989) (Tab L Research Report
Summary #27). (We were unable to obtain this Report.) There were, however, problems with this
test and as a result new test firings are being conducted.

Because PNNL had conducted many of the earlier tests, and had developed health and safety
guidelines for DU munitions, CHPPM requested PNNL to perform detailed reviews and analysis of
data from burn tests of DU munitions to assess the possible chemical and radiation exposures to
personnel at Camp Doha. PNNL estimated that 660 rounds were involved, totalling 3090 kg DU. Of
this they estimated that 408 kg DU oxide was produced, and that 0.1% was a conservative (upper)
estimate of the fraction released to air. For resuspension, they took the average surface
concentration (0.4 g m–3), and resuspension factors (concentration in air / concentration on ground)
of 10–4, 10 –3 and 10–5 m–1, before, during and after decontamination, respectively. Estimated intakes
are not given but maximum kidney concentrations were all below 0.1 µg per gram kidney, and
effective doses below 0.1 rem (1 mSv).

E4.23 TAB P – DoD and VA Medical Surveillance Programs for Gulf War Veterans (IV).
Follow-up since 1993 on 33 soldiers who were manning vehicles struck by DU munitions: 15 are
known to have retained fragments. Others might have fragments, or have DU intake by inhalation
or wound contamination. There is concern that DU dissolved from fragments might lead to toxic
effects in kidney or other organs. In 1993/4 24-hour urine samples were measured (Hooper et al
1999). The range was 0.003 – 22.48 micrograms of uranium (µg U) per gram creatinine; mean
levels were 4.47 and 0.03 µg U per gram creatinine in soldiers with and without identified
fragments. Latter are similar to mean levels in non-exposed population. Levels measured in 1995
were similar.

In 1997, 27 of the 33, and 38 veterans not exposed to DU, were examined (McDiarmid et al 2000).
Uranium levels in urine were similar to those previously found. In the DU-exposed group, urine
concentrations ranged from 0.01 to 30.74 µg U per gram of creatinine. In the non-exposed group,
the range was 0.006 – 0.047, which generally agreed with concentrations in previous studies of
unexposed populations. Gamma-ray spectrometry (whole-body counting) was also used to measure
the radioactivity in the body. In nine veterans, this indicated the presence of uranium above the
limit of detection. The results correlated well with the urine tests. However, the urine tests are
considered more sensitive in relation to kidney risk, since the spectrometry did not detect uranium
in some individuals who have elevated urinary levels. The estimated annual effective dose produced
by the uranium released by the embedded DU fragments for the nine individuals is below the
estimated background exposure to the general population of 0.36 rem (3.6 mSv: note that this is
higher than the estimated UK average of about 2.2 mSv) and only one individual is above 0.1 rem
per year (1 mSv).

In 1999, some participants were re-evaluated and 30 new friendly-fire victims were added to the
programme. Four have identified DU fragments and “elevated” U in urine (above 0.05 µg U per

19
gram of creatinine). Evaluation of the 1999 data is ongoing.

A separate follow-up programme was initiated in 1992 to evaluate the exposures of the 144 th
Service and Supply Company, the unit that operated the damaged equipment yard at King Khalid
Military City. This is discussed in annexe C, section C3.6.

In 1998, the DU medical follow-up program was expanded to evaluate veterans who received the
largest DU exposures during the Gulf War and to ensure that veterans with higher-than-normal
levels of uranium in their bodies are identified and given appropriate medical monitoring. This
programme requires participants to submit a 24-hour urine sample to establish a urinary uranium
level and a detailed DU-exposure questionnaire. The notification and medical evaluation aspects of
the program are described.

OSAGWI has been identifying Level I and II veterans, and inviting them to participate. Initial focus
was on Level I. All 104 survivors who were in or on vehicles that were struck (including the 33
already in the VA programme) were identified, and all but a few contacted. OSAGWI is continuing
to identify soldiers who entered burning DU-contaminated US vehicles immediately after impact. To
date, 45 have been identified and 36 of them notified. OSAGWI has also so far identified 127 Level
II veterans and notified 117. (Eight have declined to participate, leaving a total Level I and II named
for follow-up of 211.)

This programme was also opened to any Gulf War veterans who were not targeted as Level I or
Level II but who are concerned about possible DU exposure. So far almost 400 have taken up the
offer. A 24-hour urine sample is taken and measured “at a single laboratory”. Recommendations
for follow-up depend on whether the urinary uranium level is normal or elevated.

Emerging results
(The reference for these is an internal document: Baltimore Division, VA Maryland Healthcare
System, "Status Report–Outpatient Urine Uranium Screening," October 2, 2000. We do not know
the technique used to measure uranium in urine. In particular, does it determine isotopic
composition?) As of 30 September 2000, 662 individuals were identified for the DU follow-up
programme, 606 since the programme started in August 1998. The programme has sent out 550
urine collection kits and received back 322 specimens. To date, 11 individuals have initially tested
above the screening guideline of 50 nanograms of uranium per gram of creatinine (ng U/g
creatinine). Seven of these have submitted follow-up samples and three of the follow-up samples
were also elevated. Four individuals have either not resubmitted samples for analysis or their results
are pending.

A total of 126 individuals in Levels I and II have either been brought to Baltimore for an evaluation
or been mailed a kit for the 24-hour urine sample.

A total of 47 DU-exposed individuals completed evaluations between 1994 and 1999, including 21
in Level I and 26 in Level II. Out of these, there are urinary uranium test results for 15. One veteran
(wounded in a friendly fire incident) had slightly elevated levels. The other 14 were within normal
limits (less than 50 ng U/gram creatinine).

E4.24 TAB Q – General Accounting Office Comments

(This report (GAO 2000) was obtained, but it is convenient to cover it here as it was summarised,
and responses to it addressed.)

20
Because of concerns about the use of DU in past, current and future military engagements,
Congress directed the General Accounting Office (GAO) to address three specific issues:
• What is the scientific understanding about the health effects from exposure to depleted
uranium?
• Are Gulf War veterans experiencing administrative problems with the current medical screening
program for depleted uranium health effects?
• To what extent have the services implemented programs to train service members to safely
operate in a depleted uranium-contaminated battlefield?

In its report (GAO 2000) GAO responded to each of these issues. With respect to health effects,
GAO cited the recent studies by RAND (Harley et al 2000) and by the Agency for Toxic Substances
and Disease Registry (ATSDR 2000), which concluded that “current evidence suggests that it is
unlikely that inhaled or ingested depleted uranium poses a radiation health hazard, namely
cancer”. GAO noted that both reviews cited the kidney as the organ that would show the first
adverse health effects from DU’s chemical toxicity, and that animal studies had shown that very
high doses of uranium may cause kidney failure. GAO (2000) also cited the findings of the DU
Medical Follow-up Program, (Tab P), which indicated that no kidney damage was found in the
participating veterans.

While GAO acknowledged that the military services had developed DU safety training, they cited
problems with verification. In response, letters were sent to the heads of the services, reiterating
the requirements to ensure compliance with the directives on DU training. A separate letter was
sent with regard to the specific training status of soldiers in Kosovo.

GAO also noted concerns about the reliability of CHPPM’s dose estimates for Level I veterans. A
major factor was the actual time for which air samplers ran in the key impact tests (Fliszar et al
1989). (GAO (2000) refers to several sources of uncertainty, including assumed breathing rates, and
types of armour.) Hence, OSAGWI directed the US Army to undertake tests to characterise fully DU
and its health and safety aspects in combat vehicles struck by DU munitions. This testing is
expected to be completed in fiscal year 2002.

E4.25 TAB R – Changes in this Report

E4.26 TAB S – Bibliography

E4.27 END NOTES

E5 Harley et al (1999)

A Review of the Scientific Literature as it Pertains to Gulf War Illnesses. Volume 7, Depleted
Uranium. RAND Corporation National Defe nse Research Institute. Washington, USA.

This is a comprehensive review (~100 pages) of the literature that is mainly concerned with the
health effects (radiological and chemical) of uranium. Nevertheless it considers exposure pathways
(internal and exter nal) and thus has some information relevant to this assessment. In this respect it
makes frequent references to AEPI (1995) and CHPPM (1998). CHPPM (1998), a draft report, was
not generally available, but a later version, CHPPM (2000) is available on the I nternet and the same
web site as this review.

21
E5.1 Introduction
The following relevant summary is included (page 5):
“The pyrophoric nature of DU is of special relevance to the health effects resulting from DU’s use in
munitions and armor. Both the impact of a DU penetrator on a target and the burning of DU
produce DU dusts or aerosol particles. In addition to resulting in aerosol particles, when DU burns,
the high temperatures created act to oxidize uranium metal to a series of complex oxides,
predominantly depleted triuranium octaoxide (U3O8), but also depleted uranium dioxide (UO 2), and
depleted uranium trioxide (UO3) (AEPI 1995, CHPPM 1998). Upon weathering, the nonoxidized
small particles and surfaces of remaining uranium metal will also slowly oxidize to those three DU
oxides over time (CHPPM 1998).

Originally it was thought that up to 70 percent of the DU round may be aerosolized upon impact of
a DU penetrator on its target or in fires in which DU burns (AEPI 1995). However, based on more
refined testing, the percentage of the original material to aerosolize is now known to range from
10 to 35 percent with a maximum of 70 percent (CHPPM 1998). The percentage varies according
to a number of factors, such as the hardness of the target, velocity and angle of impact, pathway
through the target (ie what it impacts--engine, DU armor, etc). If the round easily penetrates a
target, as it does non-DU armor, less of it will aerosolize than it does when hitting a nonpenetrable
target, such as laminated steel. In many cases of DU hits during the Gulf War, the DU penetrator
went completely through (out the other side) the target armor.

Although other industrial compounds exist, the uranium compounds of concern from the military
use of DU are primarily limited to the uranium oxides. In the body, the oxides are mostly
metabolized to the uranyl ion (UO2++). This discussion is limited to the uranium compounds present
in a military environment, namely DU metal and its oxides: U3O8, UO2, and UO3.”

(We found no reports that UO 3 was formed immediately in impacts or fires. CHPPM (2000, P J-90)
states that “Only DUO 2 and DU3O8 have been observed from either hard-target tests or fires.
However, because DUO 3 has been found in some of the range studies and is considered to be the
most soluble of the three DU oxides, it was chosen to model the soluble fraction.”)

E5.2 Health Effects

The main part of the report (46 pages) includes in the context of inhalation (page 12): “The
percentage of DU aerosol particles that are respirabl e varies according to the circumstance by
which it oxidizes. For particles generated by fire, the percentage smaller than 10 µm AED ranges
from 0.1 to 33, while particles generated from impact of a hard target are virtually all smaller than
10 µm AED (CHPPM 1998). It is the respirable particles that present a potential health hazard from
uranium inhalation…

It is unlikely that any munitions explosion involving DU could have sustained air concentrations of
DU in the mg/m3 range (1,000 µg/m 3) for any length of time. Outside of struck vehicles dispersion
of airborne material by normal wind speeds and ground deposition (fallout) dilute any clouds of
material rapidly. (See OSAGWI 1998 and CHPPM 1998 for discussions of exposure.)”

E5.3 Appendix G Measured deep dose rates for M60A3 and M1 Tanks
Tables are given of measurements attributed to Parkhurst (1991). Dose rates are in the range 0.01–
0.2 mrem hr –1 (0.1–2 µSv hr–1) in the crew positions. Hence 1000 hours would give doses of 0.1–2
mSv, of the order of annual natural background radiation.

22
E6 Mishima et al (1985)

Potential Behavior of Depleted Uranium Penetrators under Shipping and Bulk Storage Accident
Conditions. Pacific Northwest Laboratory. PNL-5415. Richland, WA.

The objectives were: (i) to characterise DU oxide samples from a heat test for particle size,
morphology, and lung fluid solubility (described in annexe H) (ii) to conduct a literature search
(1979 onwards) on uranium oxidation rates, characteristics of oxides generated in fire, airborne
releas es and radiological toxicological hazards from inhaled uranium oxides. There was a need to
formulate conservative but realistic accident scenarios. Main factors: amount oxidised and particle
characteristics.

E6.1 Accident scenarios

Two were considered, wi th input from literature search on fires:
• truck with 512 rounds collides with pile of wood etc. Combustible materials, plus fuel, lead to
intense fire: 800–1000ºC, for 6–10 hours. Without action would smoulder for ~2 days;
• storage facility: ‘igloo’ with up to 3408 rounds or ‘Stradley magazine’ with up to 6176 rounds.
Both covered by earth on six sides. Combustible materials: normally only packaging, wooden
pallets, propellant. Postulated vehicle collision with entrance, provides fuel, and an enlarged
opening. Without action ‘staballoy’ (DU) penetrators subject to 300–1000ºC (temperature of
glowing ‘char’) for many days.

E6.2 Literature search

Accident conditions: fire important to oxidise and disperse oxide; detailed consideration of wood
fires.

Uranium compounds and reaction rates

Uranium oxidation usually proceeds at two rates with a “break weight or time”. Note that
staballoy is not pure uranium, but contains 0.75% titanium, which may affect oxidation. Oxidation
rate increases with time and generally produces heat. Oxide layer controls rate at which oxidant
reaches metal surface. Various factors influence oxidation rate (table 4.1). Ignition temperature
depends on specific surface area (fig 4.1). Based on experiments (including Elder and Tinkle 1980)
ignition of penetrators is not expected under fire conditions. Oxidation rate has been studied under
wet and dry conditions and temperatures up to 1400ºC. Below 300ºC it is slow, and termed
“corrosion”, accelerated by water vapour. At higher temperatures oxidation rate increases with
temperatures. Swelling and thermal stress accelerate oxidation.

Characteristics of DU oxides generated by fires

At low temperatures (<200ºC) hyperstoichiometric UO 2+x, where x is up to 2.5, forms, plus other
compounds: U3O7, U3O8, U4O7-y. UO3.0.8H2O is formed at low temperature and high humidity.

Above 300ºC, U 3O8 with hyperstoichiometric UO 2 forms (figure 4.7). Notes that the finding of Elder
and Tinkle (1980) that UO 2 content increased with temperature may be an artefact of the
experiment since the metal was not completely oxidised.

Particle size of oxides formed in a fire

Megaw et al (1961) reported size distributions of oxide produced from oxidation of metal in air or
CO2 at 600–1000ºC, at various degrees of turbulence (Reynolds number). Median was ~10 µm at
600ºC and >400 µm at 1000ºC. Refers to Elder and Tinkle (1980) and to a UK MOD “Guidance
Notes for the Storage and Transport of Depleted Uranium Munitions (4 th draft)” which indicated

23
~0.15% <10 µm dae .

Airborne release for accident scenarios

Considered the fraction <10 µm dae as potentially respirable. Several events/mechanisms are
possible for release during or after formation of oxide: free-fall spills, combustion of substrate onto
which powder lands, flaring, explosion, resuspension. Table 4.7 reviews published values: there is a
very wide range 0.00003–0.3. Resuspension factors are also variable (page 4.34): “Resuspension
factors have been found to vary over 11 to 13 orders of magnitude”. Computer codes have been
developed to predict airborne releases for some accident conditions. Concludes that for truck
accident, <0.1% of the U oxidised would be released as respirable particles. Bulk storage accident
much more difficult to assess.

Downwind transport of airborne DU particles

Atmospheric transport models treat material as a ‘plume’ (continuous release) or ‘puff’ (short-term).
Considers each factor in turn:
• source term: description of material released to atmosphere; activity of each radionuclide; mass
of toxic chemicals;
• transfer from release to plume: could be 100% for open fire or explosion;
• atmospheric dispersion model accounts for initial size of plume, entrainment of air and/or
resuspended material, building wake, plume meandering, plume rise due to momentum and/or
buoyancy. Many difficulties in predicting near-field concentrations (< 1 km);
• effects of complex terrain and rough surfaces;
• calculation of downwind atmospheric transport. Gaussian model most commonly used.
Describes plume concentration in terms of horizontal and vertical standard deviations, which
depend on distance, wind speed and atmospheric stability;
• end points are air concentration (for inhalation), and via deposition velocity, ground deposition.

Potential inhalation hazard from airborne releases of DU

Recognises radiological hazard mainly to respiratory tract from insoluble particles and chemical
toxicity mainly to kidney from soluble material.

Describes ICRP publication 30 lung model and variation of doses with particle size. Notes that (page
4.56) particles <10 nm may pass directly into blood. Notes that test indicates <0.1 wt% of U
released, but for worst case assume 4% released as respirable particles. Discusses evidence for
chemical toxicity and various exposure limits:
• 0.2 mg m-3 TLV (Threshold limit value for continuing, repeated exposure, ACGIH 1983);
• 0.6 mg m-3 STEL (Short-term exposure limit value for 15 minutes, 4 times a day, ACGIH, 1983);
• 30 mg m-3 IDLH (immediately dangerous to life and health, maximum level to escape within 30
minutes without permanent injury, NIOSH 1978).

These are used to calculate an ‘exclusion area’ following an accident. Notes that exposure limits are
generally based on continuous exposure, but after an accident, exposure is short-term. Derives
appropriate limits from maximum permissible concentration in kidney of 3 µg U per g tissue. Gives
25–40 mg-hr/m3. Brief consideration given to other exposure pathways: external radiation,
ingestion following deposition on plants and contamination of water, resuspension of deposited
particles, but dismissed as insignificant compared to inhalation.

24