NOx & Ozone Gas Phase Titration (GP)Calibration (August 2008, rk)
The Gas phase titration method to calibrate NO, NOx, NO2 and O3 is based on the one
to one reaction of NO +O3 to form NO2.
NO + O3  NO2
Things to turn on:
1. Check that NOx and O3 meters are on; usually they are on all the time. If thye
are not, tune on the computer remote switch “S outlet mixing fan”
2. Check that the GPC air tank is on in the utility area. If it is empty you will have to
use the back up pump on the wall (Figure 1)
3. Remote switch “S inlet mixing fan” shold be on. This turns on the solenoid on-off
valve on the GPT air tank in the utility area, and the Tylan flow digital meters. It
also switches the NOx and O3 meters to the GPT cal flask from the Mahogany
sampling board.
4. When finished turn off the remote switch on data system, set to South inlet
mixing fan to off). Doing this will return the NOx and O3 sampling lines to the
chamber gas sampling manifold behind the particle instruments. Make sure the
NO cal tank is off.
The general Idea: The NO instrument is 1st calibrated by flowing and diluting known
amounts of NO from a calibrated tank (usually at ~50 ppm + 2 %) with NOx free, clean
air, into a flask. This produces a known concentration of NO at the outlet of the flask.
A NOx instrument and an O3 instrument ”look” at these known concentrations. NOx free
air usually flows into the flask, that will be called a GPT flask, at a flow of 2.5 to 4 L/min.
(See Figures 1 and 2 below). C8urrently we are using a Tylan flow of 3.5 l/min.
O3 generator
Tylan flow
meter
purifil
GPT flask
50 ppm NO
Cal tank
dryrite
rotameter
3-way valve
on/off
valve
10 L/min ss
metabelows pump
GPT tank air
“south inlet
mixing fan relay”
UV O3 calibrator or
nebulizer particle size
calibrator
Figure 1 Schematic of GPT cal system
Figure 2. Picture of GPT Cal system
In a complete calibration, four different NO concentrations are added to the flask,
starting with ~0.1 ppm and going to ~0.6 to 0.8 ppm. To do this, you will start with an
NOx flow of 0.01 l/min and an air flow of ~4.00 l/min (~0.12 ppm NO). After a stable
reading, you will then increase the NO flow to 0.02 l/min then 0.04, and finally to 0.06
(~0.74 ppm). We will then add O3 to the flask by turning on the inline O3 generator to
titrate some NO and create NO2 so that there is a mixture of NO and NO2 in the GPT
flask. Typically, NO is in excess over O3, so that not all of the NO reacts. This change in
NO gives the amount of NO2 that forms and the amount of O3 that had to react to give
the change in NO. Not all of the O3 will react in the flask, and the amount of ozone that
makes in through the flask is recorded. The NO flow to the flask is then removed and
the total O3, (un-reacted and reacted) is observed. The un-reacted O3 is subtracted from
the total O3 and this equals the calibrated change in NO.
1. To begin: The NOx and O3 instruments in the new aerosol lab should be on for at
least 2 hours. Make sure the NOx meter pump is on. The old NOx meter vacuum
should be ~ 26 inches of Hg. There is a filter in back of the NOx meter and replacement
of the filter is necessary after every 3-4 experiments. If you are using the new Monitor Labs
NOx meter (currently it is broken), you need to check the dryer on the top of the NOx instrument. If it does
not have 1 inch of blue zone, it needs to be replaced (pink indicates that the dryer is spent).
We are currently using an old Bendix Nox meter (Figure 3). The vacuum pump should
read at least 25”Hg. (Figure 4). Note, there is a back up Monitor Lab Nox meter in the utility room.
There are two components to this NOx meter. They are in separate blue instrument boxes that are on the
other side of the wall from the GPT cal area. The one on top is
the pneumatics and pump box and the other is the electronics
box. The directions on the meter say to power up the pneumatic
box 5 minutes after the main power is turned on. I don’t know why
, but this message is left over from the 1980s at the site. The
power cord usually is plugged in. If you have to use this NOx
meter, there is a sampling line that run into the For the old Bendix
Nox meter that we are currently using, make sure the put the NO
and NOx settings on the range scale of 1 ppm. Do not move the
span or zero settings. If using the back up Monitor labs Nox
meter, on the pneumatic box on top, see that the O 3 rotameter is
30-40% of full scale and the other two are ~80% of full scale. (If
you add more than 1ppm of NOx, change the NOx & NO2 to
Figure 3. Old Bendix
2ppm. Do this for both instruments)
Chemilumenescent Nox meter
2. Bring up the NOx computer (usually it is on, but the
monitor may be off; it is next to the black data logger).
The Tylan flow controller boxes (see Figure 2) should be on, because they are turned on
by remote switch S inlet mixing fan. If you do not see the digital numbers on the Tylan
flow meters, turn on remote switch “South Inlet Mixing Fan. This also switches the
sampling of the NOx and O3 instruments to the
GPT cal flask form the Mahogany board chamber
sampling manifold.
3. The general procedure to calibrate the NOx
meter will be to first take zero NOx and O3
readings with just NOx free air flowing through the
flask; then you will create dilute mixtures of NO in
nitrogen in the GPT dilution flask. NOx free air to
the GPT flask can come from an air tank in the
utility area called GPT air, or if the tank is empty,
Figure 4. Pressure Gauge on Old
from a pump on the wall. Note the air goes
Bendix Chemilumenescent NOx
through a dryer column (should has at least 5 cm
meter
of blue indicator) and a purifil column to remove
NOx. Adjust the air flow to 3.5 liters/min on the Tylan air flow controller read out, by
adjusting the air flow rotameter in Figure 2. The Tylans read in liters/min at zero oC.
Beginning the calibration:
5. It is very important to purge the NO regulator before conducting a cal: This
procedure is performed so that any air in the regulator that results during no operation,
does not back diffuse in the NO tank. If it does, O2 will slowly oxidize NO and ruin the
tank. Typically, you may find the NO flow controller needle valve connected to the
regulator and this supplies NO to the Tylan flow meter. You will have to disconnect this
apparatus (see red male quick disconnect and the NO needle valve), since this is the
normal position you will typically find the system in (see picture above, with labeled NO
flow needle valve.). After removal replace this with another male quick disconnect that is
attached to an 1/8” tube that vents directly into the vertical waste pipe behind the NOx
instrument. To purge the regulator, close the low pressure regulator outlet valve near
the quick disconnect. Then quickly open the main valve of the tank and close it quickly.
You should see the pressure on the main tank gauge of the regulator to ~1000 psi. Now
vent the regulator by opening the regulator outlet valve. When all of the pressure on the
main regulator gauge goes to zero, close the outlet valve, and quickly open the main
tank valve. Repeat this 5-7 times. Again, this is very important because it purges
oxygen from the regulator which can back-diffuse into the tank and slowly oxidize NO to
NO2. Now adjust the tank outlet pressure to 10 to 15 psi.
7. The Tylan flow controller Flows: It would be great if we could read the Tylans
directly, apply a calibration factor and then correct them for the temperature in the GPT
flask, to get the real flow. For now we are checking our previous Tylan air flow
calibration with a “spot check flow” from a calibrated flow meter. If this is ok, there is no
need for a more rigorous bubble meter flow calibration. The Tylan air flow when
corrected to room air temperatures reads about 2-5% low.
A check on the Tylan flow meter is performed by adjusting the air flow with the air flow
rotameter valve on wall to ~2.5 to ~4 l/min. When the flow is stable on the Tylan read
out, check the flow the large 496 ml soap bubble flow meter or the “dry cal test meter”
Get 3 stable stopwatch times (say 12.36 sec+ 0.08 sec); record these and the air flow
from the Tylan flow read-out in the spread sheet cells under the graph (Tylan air flow
cal) on the left side of the spread sheet. Also enter the temperature from the
temperature sensor near the GPT flask. The spread sheet will now estimate, based on
the last cal, the predicted flow to the flask. This should agree with in 5% of the soap
bubble meter. Now remove the air flow from the bubble meter and re-attach it to the
GPT flask, adjust the air flow to ~ 3.5/min so we can get GPT zero readings with the
NOx and O3 meters. The NO inlet to the GPT flask should be capped and the NO flow
should not be attached to the GPT flask, because we will spot check its flow.
Check the NO Tylan flow the same way you did the air flow, but use the 100 ml burette
bubble flow meter. This, however, must be done for each NO flow that is used the Tylan
flow meter is not stable from day to day.
A note on the bubble meters; make sure you get good clean bubbles that are circular.
Your should be level with the bubble as it crosses the line on the bubble meter at the
bottom and at the top. Ideally, only one bubble should be in the bubble tower at a time.
Readings should agree to 0.08 sec.
8. Adding NO to the GPT flask: Bring up the NOx-cal template excel spread sheet and
give it today’s date.
Chart readings
NOx cal worksheet (change grey shaded cells only and save under a different name, blue cell values are
calculated)
OPERATORS:
DiHu, SL
DATE:
6/3/2007
Tank ID
Tank NO conc=
NO2 =
Our new tank
National specialty gases #cc51730, 2-20-98
50.5 ppm
0.82 ppm
With NOx free clean air (often called zero air) flowing to the GPT flask and the NOx and
O3 meters sampling form the rear of the flask record the the NOx and O3 GPT zero
values. Note there are two places to record zero values. There is a bank set of cells the
other says “GPT air concentrations only.” These should be the same, but if say were to
humidify the air and look at how the NOx zeros change with different humidities we
could record this in the top boxes. Typically we just record the GPT air concentrations
only.
NO
0.000
NOx
O3
0.000
0
GPT air concentrations only
NO
0.005
NOx
-0.002
O3
0
The calibration can now proceed with either a one point or a, multi-point calibration. One
point calibrations are done before each experiment. To do a 4 point calibration, 4
different NO flows have to be introduced to the to the flask (say 0.01, 0.02,0.03, and
0.04,l/min) waiting 7- 10 minutes between flows. The air flow can be left constant. Note
in the spread sheet below the is a column to enter the air flow. Over the years we have
found the Tylan air flow to be within 5% of the soap bubble meter so typically we just
take the Tylan air flow reading and correct it from STP (0 oC) to the temperate at the
GPT flask. The current version of the spread sheet reflects this. To increase the NO
flows open the needle valve slightly till you get the desired flow. Record the raw Tylan
air and NO flows and enter these into the spread sheet.
1st Calibration
air flow
NO flow
l/min
seconds
6.36
6.4
6.46
6.43
6.36
37.83
38.05
37.96
37.8
Soap
bubble
meter
volume(cc)
avg(sec.)
496
6.402
10
37.91
flow(L/min)
4.648547
0.01582696
3rd
Calibration
NO flow
seconds
2nd Calibration
airflow
NO flow
seconds
seconds
15-Jan00
15.76
15.63
15.69
15.7
496.00
seconds
L/Min
4.648547329
10
15.7
0.038217
seconds
L/Min
air flow
seconds
496
#DIV/0!
10
#DIV/0!
#DIV/0!
#DIV/0!
Again, enter the temperature at the GPT flask during each flow change. The spread
sheet will compute the correct flows for air and Tank NO based on the temperature of
the flask
In the cells below, the spread sheet calculates the NO and NOx concentrations in the
GPT flask. Record the the observed measured values in the appropriate cells.
Based on these flows the following NO
and NOx conc result from
Cal
Number
#1
[NO ppm]tank x NO flow/total flow = [NOppm]flask
predicted
concentration
NO
NOx
O3
NO
meter
readings
NOx
O3
0.732
0.718
0.720
0.744
0.004
Calibration factors
NO
measuredbkg
cal factor
#2
measuredbkg
#3
Calfactor 3
#4
Calfactor 4
The way the NOx spread sheet computes the NO concentration in the GPT flask, is to
start with a mass balance on the flask. That is: the mass/unit time of NO flowing into the
flask = the mass/time fo NO flowing out of the flask or:
[NOtank] x NO tank flow = [NOGPT] x total flowflask
Say that the Tylan corrected air flow at 27oC = 4.00 liters/min to the GPT flask and the
NO corrected flow = 0.02 L/min at 27oC, and the
NO tank concentration is 50.08 ppm
[NOGPT] = [NOtank] x NO tank flow / total flowflask= 50.08x 0.02/(4.00+0.02)= 0.249 ppm
After you are finished with the cal, enter the calibration slopes into a file called NOx cal
history. This way we can keep track of how the NOx meter is performing
Make sure your NO predicted values do not exceed 0.9 ppm. The NOx meter
overloads beyond 1 ppm. The multi-point cal spread sheet computes a regression plot
in the spread sheet and will give you NO and NOx calibration lines in y= mx+ b format.
9. Calibrating O3 to the GPT flask: Starting with a stable NO and NOx concentration in
the range of 0.90 to 0.6 ppm, record the O3 reading at the back of the flask. On a one
point cal, the NO and NOx before the titration is keyed to the observed NO and NOx
Ozone Calibration:
Titration
meter
NOx
0.720
NO
0.718
before
O3-back
of flask
 NO after
0.323
adding O3
Corrected NO
NO xcal factor
O3
-0.0022
0.707
Corrected
NO
0.407
0.075
NOx
The reading for O3 should be equivalent to an O3 zero. Then turn on the O3 generator.
This is the switch on the black box above the GPT flask on the wall (see picture). NO
should decrease and NOx should more or less stay the same. Wait until NO and NOx
are stable. This may take 10 minutes (because the UV lamp needs to warm up).
Remember NOx should stay stable, even though NO and NO2 are changing. If NOx
changes by more than 5% there may be a problem with the NOx catalytic converter. If
we got a good NO and NOx cal first, even if NOx changes slightly due to flow changes,
we can get back to the numbers. The assumption here is that the catalyst is working
correctly. It is important to be taking O3 data at this point, because we need to assess
the amount of O3 which does not react in the flask and makes it to the outlet of the flask
before reaction (O3 flask outlet) .
When NO and NO2 look reasonably stable after adding O3 to the GPT flask, record
these values. A delta NO is calculated next, by subtracting uncorrected NO both before
and after O3 is added. The correct NO is estimated by multiplying by the NO cal factor
(slope).
10. After NO2 and NO are stable remove the NO flow from the flask and shut off the NO
tank, this is very important. Also quickly cap the GPT flask and take O3 data. You will
need to take at least 6-8 readings. Record the raw O3 value.
remove NO flow to flask
predicted O3= corrected delta NO
corrected
delta
NO
0.407
raw
O3 meter
0.409
Corrected
meter O3
O3-O3fsk
0.334
NO
flow
factor
1.0166
O3 cal factor
correct NO/O3=O3flask
*NO flask factor
1.238744
Recall that the amount of O3 which made it through the flask is entered in or near cell
D59. This will be subtracted from the total O3 measured when the NO tank flow is
removed from the flask, and this difference is equal to the corrected NO. The corrected
 NO conc. is repeated in O3 calibration calculation above. Lastly, when we removed
NO from the flask, the total flow to the flask flow decreased by a small amount equal to
the NO tank flow. This must is taken into account, since the NO tank flow is no longer
diluting the flask. It is computed by taking the original total flow and dividing it just by
the air flow. This number is entered in cell for “NO flow factor” and then factored in
computing the corrected meter O3. Finally, this corrected O3 is compared with the meter
O3 and a calibration correction factor is estimated to be 1.24 in the above calculations.
Record the O3 cal factor in the NOx cal history file. Turn off the O3 generator.
Check that the NO tank is off
10-16-96 rk
rev: 3-31-97 kl
rev 3-23-04 rk
rev aug 14, 2008, rk