Dear AMMRL,
Many thanks to all of you who provided accounts of your
experience with oxygen safety sensors in NMR labs and advice on
sensors with longer lifetimes and fewer false alarms than the
Industrial Scientific AirAware sensors we currently use. Special
thanks to Ulrike Werner-Zwanziger, who summarized a similar query in
a recent AMMRL posting. The results below are based on responses to
both her and my queries.
Based on the numerous responses, it appears that there are
basically three technologies available for measuring ambient oxygen
levels in the NMR lab:
1. Electrochemical sensors - these are by far the most common type
used in both fixed and portable oxygen monitors, including the
AirAware unit. While they all require periodic recalibration and
sensor replacement, the frequency with which recalibration and sensor
replacement are required and with which false alarms occur seems to
vary from lab to lab, with most reports describing a sensor lifetime
of at most 1 to 2 years and recalibration intervals of 6 to 12 months
or shorter.
One response suggested a possible explanation for the short
lifetime reported from electrochemical oxygen sensors in some labs:
high ambient levels of formaldehyde (from latex paint), NOx, ozone,
singlet oxygen, atomized oil (from pumps or compressors) or soot.
Perhaps these contaminants are especially abundant in the air of labs
where these sensors fail quickly. Other caveats reported related to
possible miscalibration in the presence of helium gas or failure due
to freezing by cryogens:
>Beware, most oxygen monitors use "electrochmical" sensors. They may
>not give you an accurate measurement of the Oxygen concentration in
>the
>presence of Helium. The results can be fatal!
>
>See http://www.enmet.com/specs/HeliumOxygenMonitoring.pdf.
> I realize that asphyxiation, and not cold, is the life safety risk,
>but I wonder if anyone has ever used temperature sensors to drive an
>emergency quench vent system. They would be much cheaper to install and
>have an indefinitely long service life. The other bad news about oxygen
>sensors, as the people at NANUC in Alberta found out when their 800
>quenched, is that oxygen sensors can be frozen so fast that they never
>activate the intended alarm.
In summary, many people reported dissatisfaction with the performance
of these sensors and also expressed concern about their propensity to
give false alarms while there were a few reports of long-lived,
low-maintenance electrochemical sensors. Since the vast majority of
responses dealt primarily with electrochemical sensors, I have
appended them (in redacted form) at end end of this E-mail.
2. Zirconium diode sensors
>We have just installed sensors from GfG company (model ZD21) that
>are zirconium diode based. They are supposed to last at least five
>years, only require calibration annually, and have a response time
>of 2-3 seconds. These are all factors surpassing the performance of
>the EC detectors. So far no problems, although we have only had
>them installed for a few weeks....
>We've just entered the O2 sensor world a month ago so haven't
>got much experience. We went with Zirconium Diode sensors
>from gfg. They have an expected lifetime of 5 years with
>calibrations likely once a year. I *think* the US office
>is in Ann Arbor (800-959-0329). We talked with Doug Boyce
>(304-365-0032). The web site is www.gfg.biz. We ended up
>with a GMA-41 controller with ZD21 sensors. The cost/year
>is about the same as the electro chemical, you just don't
>have to calibrate as often.
3. Optical absorption sensors
>We were concerned about the reliability of the electrochemical cell as
>well. Therefore, I was able to find a company called Oxigraf that
>manufactures an O2 sensor called the O2iM. It uses a laser to measure
>the absorbance at 760nm which is absorbed by O2. It requires two
>calibration points (we use 100% O2 and an air standard ~22%). It is very
>reliable and the service is very good. When we first received it, there
>was a problem with the firmware and the O2 concentration reading would
>dip slowly down ~19% over days. I sent it back and they repaired it
>quickly at no cost to me and its been working great ever since. I
>calibrate it maybe once every other month and it takes about 5 minutes.
>It's a bit pricey at around ~$2K but its very hassle free w/ no
>replacement parts. The laser has a very long lifetime. We have had no
>false alarms since it was installed.
I would conclude from the above that the zirconium diode and
optical absorption instruments look like very promising technologies
although there is less data available on their long-term performance
than exists for the older electrochemical technology. Over time,
these may in fact become the instruments of choice for monitoring
ambient oxygen in the NMR environment.
Finally, there was some discussion about whether oxygen
sensors, and particularly fixed oxygen sensors are needed at all in
NMR labs:
>Permanent sensors are a waste of money, IMHO. The risk of a quench is very
>low after a magnet is persistent and decreases over time. The danger of an
>N2 accident starts at the floor, and the danger from a He accident starts at
>the ceiling, so one needs 2 sensors to protect an NMR room. The greatest
>risk is during a fill, when there is a cryogen-competent person there to
>evacuate the room anyway.
>Our local engineer recommended oxygen sensors a few years ago. We
>looked into the subject and decided not to use them. They are
>contaminated by organic vapors and have a short life span in most
>laboratory environments. Both of our magnets are located in small
>rooms. During a nitrogen fill we have been able to make an oxygen
>sensor respond when it is placed on the floor directly below the
>magnet. We do open the lab doors when performing nitrogen fills and
>that eliminates the problem. The need for sensors is only important
>when handling large quantities of liquid Nitrogen such as prior to
>energizing an imaging magnet.
>
>If you make the decision to go ahead with sensors, then I would go in
>the direction of wall units with outside audio warning and a
>hardwire relay to campus security. Then I would insist on
>maintenance by physical plant to monitor performance and change the
>units and keep records. Otherwise you are personally liable in the
>event of an accident.
>We don't use them and they aren't generally necessary in most labs
>designed with higher air exchange. Anyway O2 sensors can last as
>little as 6 months.
While these points are well-taken, I would add that many of us do not
have a choice as to whether these sensors should be installed since
they are mandated by local or institutional safety regulations. For
my part, I will be looking into the zirconium diode and optical
technologies to see if they are affordable and mature technologically.
Best regards,
Ken.
Responses mainly describing electrochemical sensors:
>I now have Ultima XA about a year no problems. Not sure of the
>technology used.
>
>I had another type previously, but readings kept drifting up.
>
>Whichever one used, these things do need maintenance/calibration, so plan that
>into the cost.
>
>Of course the low cost solution is to get a noisy bird.
>
>Then place in your SOP:
>"When the chirping stops, everyone leave the room."
>It's in two parts:
>
>Gas sensor:
>http://www.msanorthamerica.com/catalog/product1512.html
>
>and
>
>Gas monitor:
>http://www.msanorthamerica.com/catalog/product1306.html
>
>The alarm is a buzzer and a flashing light.
>
>Actually there is a false alarm that occurs which really isn't false:
>when filling the magnet, if I'm too slow turning off the
>liquid N2, a cloud hits the sensor and then I'm busy
>reassuring people while frantically fanning the sensor :-)
>
>The bird is more fun...
>I have oxygen sensors from MSA. My experience is similar to yours. They
>say that they are good for 12 to 18 months, but they certainly seem to get
>flaky faster than that. The cost is about the same as yours. In our case,
>when they sound the alarm they turn on two overhead fans (one blows in and
>one blows out) that should clear the room of helium in case of a quench.
>So it doesn't depend on people leaving the room when the alarm sounds.
>
>I've resigned myself to replacing the sensors once a year and just dealing
>with the false alarms. I actually didn't replace them for about 18 months
>the first time, and I think they didn't have too many problems during the
>first year.
>I agree that it is worthwhile to address this issue again.
>Normally I take the position that safety should have top priority, and
>cost should be no object. But if the sensors fail so frequently that
>we have all learned to ignore them, then their value is questionable.
>
>My own experience is that the O2-sensors have operational lifetimes of
>a little over a year. At my previous job we had sensors installed in
>only one of our NMR rooms (the most recent addition). We had the added
>misfortune of having the O2-sensor set up to open an exhaust fan. This
>event happened on one particularly snowy weekend, and we came in to
>find a rather large puddle of slush in the NMR room.
>
>In my current position we have O2-sensors in four labs (one pair in
>each lab for a total of eight sensors). The newest installation does
>also have an exhaust fan, but the fan is manually operated from a
>switch near the operator. (The NMR magnet is located in one room while
>the operator is in another room. The two rooms are separated by a
>glass wall.)
>
>So what is my take on it? If/when I participate in designing a new NMR
>lab, I will bring in the O2-sensor manufacturers/suppliers and demand
>that they convince me that what they are selling is truly worth the
>cost. If people just ignore them when they go off, then what's the
>point of having them. And let's be honest, if the magnet quenches,
>there are alot of telltale signs it's happening.
>We use a wall mounted oxygen sensor that I have been very happy with. It
>has a replaceable sensor, but it has been working fine now for over 5
>years, without needing replacement. I do have to calibrate it every 3
>months or so, but I think all sensors need calibration. It has a built in
>alarm, plus a relay switch which I have connected to an external light and
>alarm system.
>
>The price for the unit was $432, and a replacement sensor is $198. The
>sensor is supposed to have a 10 year life, and as I said, so far I have not
>needed to replace the sensor.
>
>
>The company is called Quartrosense Environmental LTD (QEL). I use the
>Oxygen Sensor model QAS-120. More information can be found here:
>
>http://www.qel.dedesco.com/part_number.html
>We have relied upon a Crowcon II oxygen level sensor for several years. The
>unit is a hand held model with a loud audible warning sound and a bright
>visual warning flasher. The set points are user programmable with respect
>to when to activate an alarm. Unit cost ca. $450 and has a sensor that MUST
>be replaced once each year at a cost of ca. $80.
>
>I often work alone in the lab and would highly recommend this device as a
>necessity in all NMR labs.
>Welcome withoin the club of unhappy people.
>Your description could be mine. -(
>
>I switched off the whole system, because nobody believed the false
>alerts and even
>more I have an air condition system with an air change of 5000 m**3 each hour
>(in cubit feet this could be somewhat around 50 000).
>
>But even more. During the maintenance I have permanently to fight with the
>technicians. They give no notive, at whoich time they arrive and the sensors
>are located nearby two magnets. They disturb the measurements during their
>maintenance.
>Our O2 sensors (first installed ... 1995?) also
>deteriorate quite rapidly, sending of false alarms every few months.
>Recalibrating helps for a similar period. After ca. 2 years (on average,
>differs vetween the individual sensors) further recalibration becomes
>virtually impossible. Solutions (so far):
>- de-activate specific sensor and/or
>- ignore alarms, or
>- install new sensors every 1-2 years (I can't remember what the price
> was, we did it once or twice ...)
>This is just to let you know that we've got an excellent 02 monitoring device:
>It's quite handy (not much larger than a cigarette-box), you wear it on
>your belt.
>It even works at strong magnetic fields. Battery and sensor work for a year
>and more without exchange.
>It is manufactured in Germany
>name: Micro III G203
>article # 1318070 (the iron-free version)
>O2-sensor : # 1318230 (one year) or # 1318231 (two years)
>
>for more informations see: www.gfg-mbh.com
>We have a protable and fixed one:
>
>The portable one is used in case of nitrogen fill by the staff who
>is doing the N2 filling:
>
>(1) The pocket calculator sized detector is a Gasman II,
> manufacturer: Crowcon Detection Instruments Ltd., 2 Blacklands Way,
>Abingdon Business Park, Abingdon, Oxfordshire, OX14 1DY, England,
>web page: www.crowcon.com.
>
>(2) The fixed one: I could find only labels on the housing:
>Zellweger analytics (german branch of Honeywell,
>www.honeywellanalytics.com) for gas detection systems. The detector
>itself is place above the magnet.
>Our O2 sensor was going off quite a bit, and we complained about it,
>and the problem has gone away, so I assume they changed the sensor.
>Our's was going off all the time and we turned off the alarm for a
>while. I know what you are going through and do not know of a
>better solution. Let us know if you find anything new out.
>We have a Drager Point Guard II Oxygen Sensor in the NMR lab here. It
>is permanently mounted. We have had it in operation for the past 2
>years and it has been trouble free. I know that it works because when
>we had a magnet quench here a while ago it went off. The oxygen level
>that it reads varies by no more than 0.2 from one day to the next. I
>don't know how much it costs because it was part of the room renovations
>when we moved in here. Hope this helps.
>When we moved the bruker system we also moved a Alpha Omega Series 1000
>Oxygen Monitor which uses a weak acid electrolyte system rather than the
>KOH electrochemical cell. It was running over more than a year before
>we move it. I checked with the other scientist and he indicated that
>the system worked well and they did not have any false alarms. He
>mentioned that during LN2 fills it used to show low oxygen (< 19.5%) but
>after that it will be normal.
>
>The company is www.aoi-corp.com (800 - 262 - 5977)
>
>I have not installed this in our lab (we will be doing it soon).
>We use a sealed factory calibrated unit. They last longer than 3 months,
>but no more than 2 years in our experience, and cost > $2000 to replace
>when they fail. I was hoping there was an alternative sensor with a much
>cheaper replacement cell that can be field calibrated. But it sounds like
>that's what you have, and you're not enjoying yours any more than I enjoy
>mine.
>We use shirt-pocket portable sensors worn by whoever is doing a fill. They
>require calibration due to battery voltage or temperature changes, and when
>the batteries are fresh the calibration lasts several days. They are turned
>off when not being worn. The sensor element needs replaced yearly and is
>reasonably priced. They are RKI instruments model OX-01.
>
>Yes, Oxygen sensor are high maintenance. We have 2 which are
>checked and calibrated every 4 months and replaced every 1.5 to 2
>years. Calibration is always necessary at least twice a year I
>don't think it matters what brand you have. I'm pretty sure most of
>the actual sensors are manufactured by the same company in the UK
>(CiTicel) and BW technologies or Crowcon simply use those OEM
>sensors.
Received on Mon Feb 27 2006 - 13:53:54 MST