SOP 13. Standard Operating Procedure for Ozone Preparation
I. General Statement of Coverage
Generation and storage of ozone adsorbed on silica gel at dry ice temperature is relatively safe
provided certain precautions are taken.
II. Hazard Assessment
A. Explosion - Ozone: If dry ice coolant is not regularly refilled (12 hr) or if too much ozone is
collected (> 100g), an explosion can occur.
B. Tissue Damage - Ozone: Tissue can be damaged if proper ventilation is not provided.
C. Tissue Damage - Dry Ice: Direct contact with dry ice and dry ice slurries can cause tissue
D. Ozone is an Acutely Toxic Gas (see SOP 11).
A. Available Training
See Laboratory Supervisor for specific training
B. Text and Literature References
Ozone, by M. Horvath, L. Bilitzky, and J. Huttner (Elsevier, New York, 1985)
C. CHP Appendix III (Section 23.3) Chemical Information Tables
Table 14. Chemicals With High Acute Toxicity
D. CHP: CHP Appendix III (Section 23.3) Chemical Information Table 18 lists the 8 hour time
weighted average (TWA) permissible exposure limit (PEL) for ozone is 0.1 ppm. See SOP 11.
IV. Chemical Storage
A. Special Storage
The ozone (adsorbed on the silica gel) is stored between uses in a sealed flask which is
kept cooled with dry ice and placed behind a blast shield, preferably in the designated
hood. Dry ice coolant must be renewed at least every 12 hours.
B. Securing Gas Cylinder
The gas cylinder connected to the ozonator is secured with a chain to the bench-top.
V. Personal Protective and Emergency Equipment
A. Eye and Face Protection
Safety Glasses and face shields must always be worn whenever handling the flask
containing adsorbed ozone. Safety Glasses and face shields must also be worn whenever
refilling the dry ice.
Appropriate thermal gloves or tongs should be used whenever handling dry ice.
C. Protective Clothing
D. Hearing Protection
F. Eye Wash
G. Safety Showers
A. Designated Areas
The fume hood used for ozone generation and storage and the area adjacent to vacuum
chamber will be posted as designated area for Acutely Hazardous Materials when ozone
is in use.
B. Chemical Fume Hoods
Required for ozone generation and storage.
C. Glove Boxes
D. Safety Shielding
Blast shields in the hood and surrounding the flask when connected to the vacuum
chamber should be in place at all times.
E. Special Ventilation
F. Vacuum Protection
G. Signs and Labels
1. Doorways: Acutely Hazardous Material
2. Containers: All hazardous materials must be clearly labeled with the correct
I. Fire Protection
VII. Specific Procedures
Ozone (O3) is a very unstable, explosive, and unpredictable substance. Ozone may be
collected on silica gel cooled to dry ice temperatures with comparative safety. Liquid ozone,
by comparison, is a much more dangerous substance which has an extremely explosive
tendency. Explosions have occurred with adsorbed ozone in this lab, but those explosions are
strongly believed to have been caused by rapid pressure increases, rather than any chemical
instability, as is thought to be a significant problem with liquid ozone. Since there is always
some risk of explosion, appropriate safety precautions, such as wearing safety glasses and face
shields and the appropriate use of blast shields, must always be followed.
Ozone strongly promotes oxidation, both to flammable materials and to human tissues. As
such, ozone should be removed from the laboratory environment via the hood. At low
concentrations ozone may smell like stale air, while at concentrations above 1 ppm it is a more
disagreeable sulfur-like odor. Do not inhale ozone as it may cause lung damage and direct
exposure to skin may cause dryness and mild irritation. Ozone inhalation can also produce
headaches, lachrimation (tearing), visual field changes, decreased pulse rate and blood
pressure, dermatitis and coughing. Because it is a strong Oxidizer, flammable materials,
including paper, are kept away from the ozone flask while any O3 is present. It is also
incompatible with rubber, alkali metals, bromine, combustible gases, oxides of nitrogen and
other similar materials, refer to the appropriate MSDS for details. At dry ice temperatures, the
vapor pressure of ozone above silica gel is about 17 Torr and oxygen will be displaced by
ozone at the surface sites of the silica gel.
The following standard operating procedure has proven to be very safe and has smoothly
provided a reliable source of ozone. Despite the fact that during the development of the
procedure several flask explosions have occurred, there have not been any flask ruptures or
explosions while using the final procedure over several months of smooth operation. If the
procedure is closely followed, with the safety shields in place, there appears to be minimal
danger to those working with the ozone or in its vicinity. Past explosions, which occurred
during the procedure development, were easily contained by the plastic coolant bucket and the
safety shields. However ozone is very unpredictable and care should be exercised at all times,
regardless of past experience. Safety glasses and face shields are always worn whenever any
work is done with the ozone containing flask. One or two blast shields are placed around the
ozone flask in order to protect the laboratory occupants from possible flying shards of glass in
the event of an explosion.
VII.2. Packing Silica Gel in Flask
Prior to any generation of ozone, the appropriate flask is packed with silica gel and glass wool.
The placement of the glass wool is critical for operational efficiency and for safety. See Figure
The glass wool is required to prevent silica gel from flowing with the gas (either oxygen (O2),
O3, or other gas) into the stems of the glass/Teflon valves during either the pump down or the
actual experimental run. Such contamination is inevitable when glass wool is not used. Any
contamination breaks the seal in the valve and potentially may permanently damage the Teflon
stem. The glass wool, when improperly placed, has made for potentially explosive risks. See
Safety Note 1.
Glass wool is placed within 1 or 2 inches of the bottom of the central glass stem of the ozone
flask-assembly. This glass wool absolutely must not extend more than an inch up into the neck
of the flask or there may be a pressure build-up if the coolant level falls below the glass wool.
The ideal placement of the glass wool is at a level even with, or below, the level of the silica
An additional piece of glass wool is placed at the junction of the side-arm with the neck of the
flask. Although the glass wool could probably be placed anywhere within the side arm, it is
usually placed at the junction as shown in Figure 1. The wads of glass wool typically should be
about 1 or 2 inches long and are packed quite loosely to allow free flow of gases.
Note that the silica gel is about an inch below the top of the spherical part of the flask. This
allows for a greater surface area of silica gel and produces more rapid reequilibration of
O3(ad) = O3(g) during experiments.
VII.3. Heating & Drying the Silica Gel
Water bonds more strongly to the silica gel than does the ozone. Prior to any generation of
ozone, any water adsorbed by the silica gel must be driven off because any water present will
prevent adsorption of ozone by the silica gel. The water is driven off by heating the silica gel
in its flask with flowing dry oxygen. If in doubt, heat the silica gel. Significant amounts of
water may be absorbed by the silica gel if it is left open to air for more than a couple hours. A
few minutes of exposure to air when silica gel is cooled to dry ice temperatures may necessitate
The silica gel and flask are heated by placing the flask in a heating mantle controlled by a small
variac set at about 40 to 50. Heating to much higher temperatures has unduly strained (or at
extreme temps. even melted) the Pyrex glass of the flask. Heating with dry oxygen or a dry
inert gas at a flow rate of about 3 SCFH (standard cubic feet / hr) is continued for at least 4 to 5
VII.4. Cooling Flask Down to Room Temperature in Heating Mantle
At the end of the heating period, the variac is turned off (or when possible, gradually turned
down from 50 to 0 over about an hour). The flask must remain in the heating mantle for at
least 3 or 4 hours, preferably overnight with a continuous flow of dry gas to prevent water
adsorption by the silica gel. The flow rate is usually set at about 0.5 SCFH to conserve the gas
supply. See Safety Note 2.
VII.5. Cooling Flask Down to Dry Ice Temperature
Only after cooling to room temperature for at least three hours, the flask is removed from the
heating mantle and placed in the double walled plastic insulating bucket. Pyrex can crack if
strained too much. Isopropyl alcohol is added to the bucket to a level half way between the top
of the spherical part of the flask and the top of the bucket. This flask and bucket must be
placed in the hood behind an additional blast shield.
Extra dry oxygen is set to flow over the silica gel at a rate of about 1 SCFH with a regulator
pressure of 5 to 7 psig. Finger sized or smaller pieces of dry ice are added, using appropriate
gloves and/or tongs, to the isopropyl alcohol 2 or 3 at a time until the liquid stops bubbling
vigorously. Then 10 or so pieces are added at once. Once the liquid stops bubbling, finely
pounded dry ice is added to the isopropyl alcohol until the bucket is full to its top with this dry
ice isopropyl alcohol slurry. See the Safety Note 3 regarding the preference for isopropyl
alcohol over acetone.
VII.6. Turn on the Ozonator
Once the bucket is full with dry ice, and extra dry oxygen is flowing (typically a 1-2 SCFH
flow rate with a regulator pressure of 5-8 psig for maximum efficiency of ozone generation per
amount of oxygen used), the ozonator may then be turned on. See Safety Note 4. The top of
the bucket should be covered with a styrofoam lid which has been cut to snugly fit around the
neck of the flask and between the inner edges of the insulating bucket.
If one needs to know whether any ozone is being generated, its smell is unmistakable. It smells
like very stale air. It may take a few hours or more for the stale smell to accumulate within the
hood, but once it does, the stale smell will linger for several minutes on clothing to which it is
exposed. Such mild exposure to ozone is not a concern, provided contact is minimized as much
as is practical. Avoid inhalation at the source.
VII.7. Generate Ozone for at Least 24 hours.
Typically the ozonator is run for 48 hours or more, adding dry ice to the coolant bath at least
every 12 hours, with an oxygen flow rate of 1.5 SCFH to generate enough ozone to run
experiments with ample pressure. If one wishes to run experiments within 24 hours, the dry
gas flow should be increased to at least 3-4 SCFH.
Each fill of dry ice should be large enough to last 2 or 3 hours past the 12 hour period, but that
extra time should be considered only an emergency safety buffer. Note that eye and face
protection must be worn during the refill of the dry ice. See the Safety Notes 5 and 6 for
operational and safety details. Do not exceed 72 hours at 3 SCFH of ozone generation.
VII.8. Flush Excess Oxygen with Inert Gas
After enough ozone has been generated, on the day of the experiments, argon, nitrogen, or
some other inert gas should be flowed over the silica gel for at least 10 minutes at a minimum
of 3 SCFH. The inert gas sweeps away the vast majority of the oxygen in the flask. This step
is necessary to insure the purity of the reactants in the reaction of interest. The valve is then
closed and the end of the side-arm is closed with a Teflon plug. Note that eye and face
protection must be worn. The flask is then connected to the vacuum chamber. At this point, a
blast shield must immediately be placed around the flask.
VII.9. Pump off Excess Gas
Next, the excess gas is pumped off. The gas removal has been accomplished through a few
methods. Do not use liquid nitrogen during the pump off. See Safety Note 7. The safest,
quickest, and most reliable method involves pumping off the majority of the excess gas with
the small mechanical pump connected to the ozone inlet (see Figure 1).
Great care needs to be exercised so as not to pump off the ozone. Pumping off is generally
completed within 3 minutes of pumping with the small pump is generally sufficient. One
must check the fore-line gauge every 30 seconds or so. Once the pressure falls to about 1.00
Torr (1000 mTorr on the gauge) pumping via the small mechanical pump is stopped.
At that point, pumping through the main vacuum chamber commences. Depending upon just
how far the mechanical pump brought down the flask pressure, this process has required a half
an hour to an hour or longer. The main difficulty is obtaining a more or less stable pumping
pressure. After a couple minutes, the chamber pressure might jump from about 3 mTorr to well
above 10 mTorr, overtaxing the pumps. Naturally, the flask valve must then be closed until the
chamber pressure falls back below about 2 mTorr. Pump off may be considered complete
when the valve to the chamber may be opened completely, and the pressure does not increase,
but rather continues to fall, even if only very gradually. Any further pumping would begin to
deplete the ozone.
VII.10. Running Experiments
At this point, the ozone is ready to be used in experiments. As with any of the experiments in
this (Parson's) lab, it is desirable to maintain a steady pressure. With the ozone, once every
minute or two, it is advisable to check the pressure and adjust it if necessary to keep it as stable
as possible. After about 20 to 30 minutes of data taking, it is sometimes necessary to close the
flask off from the vacuum chamber to allow the ozone desorbtion to reequilibrate. When there
is ample gas phase ozone, the chamber pressure is steadier. Typical experimental procedures
are then followed.
VII.11. Replenish Ozone for the Next Day
At the end of each day's experiments, the flask is detached from the vacuum chamber and
quickly sealed. Eye and face protection must be worn. The flask is again connected to the
ozone generator in the hood. Ozone is generated for at least 4 more hours to replenish the
ozone content of the top layer of silica gel as described in Sections VII.6 and.7. Even when not
collecting ozone, the flask must still be kept cold and the dry ice must be refilled at least every
VII.12. Release Ozone Through the Hood
At the end of the week's experiments, the remaining adsorbed ozone is released in the hood.
The valve on one end of the flask is opened and the plug on the other end is removed. The
ozone flask must be left in the dry ice slurry inside the insulating bucket.
The gradual warming of the silica gel and the release of the ozone is as dangerous as many of
the other procedures involving the ozone. Since the ends of the arms of the flask are both open
during warm up, one would think that this would have been one of the safest steps in the entire
process. However, on one occasion when the flask was removed from the dry ice slurry and
placed on the hood bench top to warm up and release the ozone, the flask exploded after a short
period despite the fact that the valve on the one end and the plug on the other end were opened.
Since that time, the standard procedure has been to leave the flask in contact with the dry ice
slurry so that the ozone would desorb more gradually and more safely. There have been no
explosions during the warm up process using this procedure.
The silica gel will need to be heated to drive off the water which will be adsorbed while the
silica gel is cold and exposed to air before it can be re-used as previously described.
Safety and General Operating Notes:
1. The presence of the glass wool, if improperly placed, can make for potentially explosive
risks. During the pump off stage prior to the experimental run, when excess gases must be
pumped out of the flask to attain a pure O3 sample, the glass wool may have trapped gases
in confined spaces above the coolant level. The rapid warm up and ozone decomposition
have led to massive and rapid pressure increases and subsequent explosions.
2. After several cycles of cooling the flask to dry ice temperatures more rapidly, the flask
eventually cracked during a subsequent use. Several hours are required during the cool-
down process to avoid straining the flask.
3. While a dry ice-acetone slurry is a more popular mixture for low temperature cooling,
isopropyl alcohol has advantages over acetone with few disadvantages. Isopropyl alcohol
has a lower vapor pressure and lasts longer in the hood before evaporating. The slurry (as
opposed to just solid dry ice) is critical to maintaining good thermal contact with the flask
and the silica gel. The lower volatility of isopropyl alcohol makes it a less dangerous
ignition hazard in the event of an explosion due to high pressure in the ozone collection
flask. Both acetone and isopropyl alcohol slurries with dry ice attain temperatures within
2C of the sublimation point of dry ice: -78C.
4. The ozonator requires a source of extra dry oxygen at a pressure of 5-30 psig with a flow
rate of 1 to 5 SCFH (standard cubic feet per hour) as measured by the flow meter on the
ozonator. Once the ozonator is properly operating, the outlet stream of gas will be about 3-
5% ozone in a stream of predominantly oxygen. The silica gel at dry ice temperature will
only adsorb the ozone. After pumping off the excess oxygen, there will be a relatively
pure ozone source remaining.
5. Although it is very difficult to use up all the ozone in one day, insufficient concentrations of
ozone adsorbed on the silica gel have caused delays in between experiments, as well as
unstable pressures during experiments. The ozone has been observed to desorb most
rapidly from the top layer of silica gel in the flask (about the top 1 or 2 cm). Sometimes,
after a week's experiments, as the ozone was released in the hood, this top layer has been
observed to be very white, with the rest of the silica gel being a deep dark purple. Purple is
indicative of the presence of adsorbed ozone. It has taken 5 minutes or more for the entire
system to reequilibrate with the top layer, and the vapor phase. The equilibrium vapor
pressure should have been about 17 Torr.
6. It has been calculated that 72 hours of 3% ozone generation flowing at about 2-3 SCFH
would yield about 80-100 g of ozone. Assuming that 100% of the ozone is adsorbed
(which I know not to be a completely accurate assumption since I have smelled ozone many
times within the hood), the yield is close to the maximum safe load of ozone on silica gel.
For about 500 ml of silica gel, the maximum safe recommended amount of adsorbed ozone
is about 120 g. Do not exceed 72 hours at 3 SCFH of ozone generation.
The maximum safe concentration of ozone on silica gel has been confirmed by other
laboratories with spark testing. Electrodes were placed within flasks containing silica gel
and ozone at dry ice temperatures. Low voltage sparks were discharged form the
electrodes. Flasks with less than the maximum safe concentrations of ozone did not
explode. Some of those with greater than the maximum safe concentration exploded. Note
that these maximum safe concentrations merely refer to chemical stability.
7. One method of pumping off excess gas has since been abandoned for safety reasons. The
entire ozone / silica gel - flask - dry ice - isopropyl alcohol combination was frozen with
liquid nitrogen. The liquid nitrogen method was abandoned after flask explosions while
warming up to dry ice temperatures. Oxygen may be freezing on some of the glass wool,
preventing a complete pump off of certain areas of the flask and, upon warming up, the
liquid or solid oxygen entered the gas phase causing an explosion. Do not use liquid
nitrogen in the pump off stage.
VII. Emergency Procedures
A. Notification: Not required unless ozone is rapidly released into the laboratory
atmosphere and high concentrations of ozone are inhaled. Contact: Principle
Investigator, University Police, and Division of Environmental Health and Safety.
B. Spill Response: In this case, a spill would actually be a rapid release of ozone to the
laboratory atmosphere. The appropriate response would be to place the silica gel (with
adsorbed ozone) in the hood and to vacate the laboratory until any perceivable odor of
ozone has been swept out by the laboratory ventilation system.
IX. Decontamination and Waste Disposal
A. Decontamination Procedures
1. Personnel: Wash skin after any direct exposure to ozone. If large amounts of ozone are
inhaled, move to fresh air and seek medical attention.
2. Area: Allow for adequate ventilation.
3. Equipment: Place equipment in the fume hood.
B. Waste Disposal:
Follow procedure in VII.12.
X. Required Approvals
Ozone generation requires the approval the laboratory supervisor. Refer to Section 20.0 and
Form 3 - Prior Approval Request.
XI. SOP Prepared by: Rodger Kampf and John Parson Date: 8/21/95
Reviewed by: Dale Karweik and John Herrington Date: 5/22/96