Hi AMMRL Community.
Here are the results of the helium recycling survey I asked for a few months
ago. Please recall that, as a followup to our discussion at ENC, I sent out
a survey on helium recycling practices, and a I got a number of very detailed
responses to that. Many thanks to Bob Berno, Brian Breczinski, Scott Burt,
John Decatur, Joe Dumais, Jeff Elena, David Jones, Sebastian Kemper, Eric
Paulson, Robert Peterson, Tara Sprules, Robin Stein, and Greg Wylie for
their help. The amount of useful information here is amazing!
At the risk of information overload, here are the compiled responses. It was
suggested that I write up a brief answer to each question that captures the
sense of current practice, but, after a number of attempts, I don't think
I can do that effectively. Practices vary significantly by site, and
conclusions drawn from one facility's solutions are often not workable at
other sites. However, respondents provided extremely valuable approaches,
insights, and tips that should be shared. The catch-all "what else do you
want to share" responses were particularly informative. I've a few
comments here and follow with the combined responses, divided into two
sections:
* one set listing everyone's responses to each question, enabling the
reader to see different approaches to the same issue
* one set listing everyone's complete responses to the entire survey,
providing a sense for situation and strategy used at every institution.
For ease of reading, I've posted a PDF of a (68-page!) Word document here:
https://www.dropbox.com/s/5xxiibuasizcvw4/AMMRL_HeliumRecylcing_SUMMARY_2022-1027.pdf?dl=0
(perhaps this could be posted to the AMMRL.org website, too).
COMMENTS:
* For purposes of this discussion, the term "recycling" refers to the
full-cycle practice of recovering helium boiloff gas, compressing it into
some kind of container, and liquefying it at a purity suitable for use in
filling NMR magnets. The term "recovery" here refers to capturing helium
gas and storing it in some kind of container like gas cylinders.
* It's complicated to determine whether recycling helium from a particular
set of NMR systems is practical.
* For a large facility with many magnets located in one room with ample
access to power and chilled water and adequate staffing, it makes sense on
many levels to install a recycling system.
* For a facility with one or two 300's or 400's with low boiloff,
problematic siting, and staffing difficulties, it probably isn't cost-effective
to purchase a full recycling system unless security of supply is especially
unstable.
* In between these situations, there is a wide range of circumstances
and support needs, making the recycle/don't-recycle choice very
site-dependent.
* The "hub-and-spoke" model: Most respondents indicated they are recovering
and recycling with a fully-connected system. But I did not hear much about more
scalable systems in which recovery happens at several locations, and recovered
helium is transported via cylinders to a central facilityfor purification and
liquefaction.
* One respondent indicated they have used such a system (in Germany)
for many years, and it works well.
* Helium consumers on my campus are pretty interested in a solution
like this, as our equipment is located in different buildings and different
city blocks.
* Costs of infrastructure improvements (piping, new electrical and perhaps
chilled water, etc.) can be major – sometimes greater than those of the
recycling equipment. It looks like granting agencies that support recycling
equipment purchases do not typically support infrastructure work, and this
must be considered when aligning support.
* This survey provided some practical ways of reducing infrastructure
costs. For instance, it has been standard for labs to use networks of copper
pipes for recovery, but some have recently found it cheaper and much easier
to use flexible corrugated stainless steel tubing.
* Whether one intends to recover helium during fills is important to determine
early. To avoid affecting a whole network of magnets, one needs extra hardware
to handle the larger volume and pressure. Recovery during fills can potentially
boost overall recovery efficiency to up to 90-95%. Though this increases initial
cost, it appears most facilities that commit to recycling include recovery during fills.
* Cost vs security: Small and medium-scale may find recycling difficult to
justify on the basis of cost alone. In such cases, it's important to stress
the security of supply, emphasizing the cost of potential quenching, potential
magnet replacement, and potential loss of research productivity due to instrument
downtime.
* The ongoing time and effort spent on helium management can be quite large, and
should be considered when deciding whether to get a recycling system.
* Some respondents estimated that helium management takes about 20% of a
facility manager's time.
* Recycling equipment appears to need a fair amount of attention. For
instance, the purification systems have traps that need to be cleaned regularly,
and this requires nontrivial staff time.
* Timing of helium "fills" is influenced by the state of the recovery systems,
like when the recovery dewar is near full. Partial helium "fills" are often the norm,
and their increased frequency necessitates increased instrument downtime and labor.
Thank you for your attention, and special thanks to the respondents for their very
detailed insights.
* Josh
P.S. As I understand it, AMMRL survey compilations have typically been anonymized to
protect privacy. Here, most respondents shared facility-specific information like
websites that describe their recycling solution, in which case anonymity could not
be preserved. When pertinent site-specific information was not shared, respondents'
names were removed from their contributions. I hope this aligns well with community standards.
Josh Kurutz, PhD
NMR Facility Manager, Chemistry Dept.
https://voices.uchicago.edu/chemnmr/
jkurutz_at_uchicago.edu<mailto:jkurutz_at_uchicago.edu>
****
SORTED RESPONSES:
0. Some responded that they do not recover helium, and they explained theirreasoning.
R0: We have not implemented helium recovery yet, largely because of cost
considerations. My facility has five instruments, 400 and 500 MHz, all with UltrashieldPlus
or Ascend magnets, so boiloff is pretty low, and we consume between 600 and 1000 L in
a given year. They are sited in two labs, in buildings separated by a city street, so it
would be cost-prohibitive to unify their helium collection. Neither NMR lab has a utility
room near the NMR instruments that would be suitable for a full recovery+liquefaction system.
We are developing a campus solution that focuses on local collection and central purification+liquefaction.
We are partnering with Biomolecular NMR and individual non-NMR labs in the Physical Sciences
Division and Molecular Engineering. We're still scoping out solutions and costs. Simply purchasing
recycling equipment will not solve our helium security problem without extensive inter-divisional
cooperation, logistics planning, and operational development, and strategic infrastructure improvement.
R2: We have chosen not to get a system since I was advised by one of the recycling system vendors
that with only a narrow bore 600 and a narrow bore 300 it was not cost effective. I personally was
concerned about the noise level since we have a small lab space and not a lot of open floor space.
My office is enclosed in a corner of the lab and between a cryoplatform and a Helium recycling system
I am concerned that the lab will be too noisy. I am also concerned about the floor space it would
consume. I am curious if these systems are able to handle power outages or if they need a UPS.
FYI we consume 100 liters of liquid Helium every 15 weeks.
R5: We don't have a helium recycling facility because, with two superconducting magnets,
it doesn't appear to be cost effective.
Q1) If you have a website describing your recovery system, please provide a link.
R3 (related): https://www.physik.fu-berlin.de/service/ttl/index.html
<
https://urldefense.com/v3/__https:/www.physik.fu-berlin.de/service/ttl/index.html__;!!BpyFHLRN4TMTrA!4ZlGqW0w1J-Iz_josZT2EfROsZnBo02PXj0AhBg9kgHp-QJYqhbqaRyeOwB2uNNitWUZD1O7R-RqJHvHHIldPTybbfNnnUZ3EdGm$>
R6: https://nmr.chem.tamu.edu/HeliumRecovery.php
R9: https://www.mcgill.ca/mc2/helium-recovery-system
R10: https://chembio.byu.edu/nmr-facility/he-recovery
R12: https://cbic.yale.edu/about-us/helium-recovery
MECHANICS
Q2) Please describe your recovery network. How many magnets? Are any of them pumped? Does your
network include non-NMR systems? Does it span different floors/buildings/blocks?
R3: 7 Magnets (200-700), none are pumped. All magnets are on ground floor
R4: 7 NMR magnets are on the network on 2 floors located directly above/below. None are pumped.
R6: 8 NMRs 400MHz and 500MHz only (number 8 being added in the Fall). 1 DNP, and 1 EPR. Oil
Pumps on the DNP and EPR. Pumps have filters on the exhaust.
Overall our system covers the entire height of our 5 story building including the basement so
6 floors total. System is divided into East end and West end they meet on the top floor and
go into the recovery room on that floor.
Tubing is all stainless steel DEF tubing (Diesel Exhaust Fluid).
R8: Our recovery system serves 8 NMR magnets (2x300MHz,2x400 MHz, 2x500 MHz, 1x700, 1x800 vintages
from Varian r2d2, oxford to Bruker Ascend), which are all located in one central laboratory. All
magnets are non-pumped. One magnet in another building is not connected to the lHe recovery system.
Also connected was a SQUID, which is a major consumer of lHe. However since COVID lab-closure this
SQUID has been shut down.
R9: Eight NMR magnets. Of these, seven (300 MHz to 500 MHz) are in one location about 100 m from
the compressor and liquefier while the eighth (a pumped 800 MHz) is in the same room as the compressor
and liquefier.
We can also collect from an EPR or attach an additional helium dewar at that point.
R10: We have 6 magnets on the system. None of them are pumped. We have 2 non-NMR magnets on the system.
Everything is in one building, but the magnets span multiple rooms across different floors (basement
through the 3rd floor).
3 magnets are in the NMR facility (300 MHz, 500 MHz, 500 MHz, solution state instruments), which is also
where the He recovery system is housed (this is in the basement at the end of the central hallway).
1 magnet is a SS-NMR magnet in a research lab (about 1/3 of the way down the central hallway, in the basement).
1 magnet is an FTICR horizontal bore magnet in a research lab (3rd floor and about halfway down the central
hallway - this magnet has the longest run of pipe to the collection bag). This magnet has much higher boil
off than the 4 NMR magnets.
1 magnet is a Quantum Design PPMS magnet. This magnet has a very high boil off. It had a cold head attached
that directly liquefied the boil off back into the magnet cryostat; however, the cold head was long past its
end of life and the performance had degraded quite significantly over the past year (the cold head was at
~80,000 hours without any servicing when it was supposed to be refurbished every 20,000 hours).
R11: 3 magnets - 1 Oxford 900 pumped magnet, 1 Oxford 600 and 1 Bruker US Magnet. All in the same room
R12: We have 10 magnets, 9 of them NMRs, one a mass spec magnet. One of the NMR systems is pumped. It's an
800 US2 from 2004.
Q3) Scale: How much liquid helium did you use annually before installing your recovery system, and how much now?
Also, please estimate the overall percentage of helium you recover.
R3: 1100 l Helium are filled annually, we recover just 50%, but we are not collecting during a fill due to a
very small balloon
R4: Annual usage before was about 2300L. It's too early to estimate annual usage now. Just installed in March.
R6: We ordered about 3600L of liquid a year before we got our recovery system. Almost 2600L were used in 2021
from the Facility. Our Facility does supply 3 Magnets that are not connected to the Recovery system. Due to
this our overall efficiency is ~85% but for magnets on the recovery system about 95%.
We currently produce about 120L a week. This does usually include at least one purchased industrial helium tank.
R8: This is difficult to say since lHe recovery was installed together with the newly purchased 700 and 800 MHz
instruments. Our total purchase volume of lHe is less than ~500l per annum. This is an estimated 20% of the lHe
we would have to buy without the recycling system (keep in mind that one magnet is not connected to the recycling
system and lHe lost during fills of the r2d2 magnet can not be recycled).
R9: We purchased about 2600 L annually before installation (April 2021) and used about 2500 L of it. In the
first year after installation, we purchased 1480 L and used 2500 L, meaning we recycled 40%. However we have
improved to 70-80% efficiency over the past few months.
R10: Ignoring the PPMS magnet, we were purchasing ~1200 L of LHe annually (8 x 60 L dewars and 7 x 100 L dewars).
It's worth noting that much of those 60 L dewars was being wasted. Only ~25 L was collected in the magnet; that
means ~8 L lost in the plume, combined with ~7 L lost in during shipping, that's ~20 L wasted x 8 times per year
~160 L that wasn't used.
The PPMS had a re-liquefier, but for the past couple years, they had to add about 1 compressed gas cylinder per
month to compensate for losses. This would be equivalent to ~100 L of LHe annually.
Before we added the PPMS to the system, I estimate that I was losing ~30 L of LHe annually (that's based on just
1 year's worth of data) and I replaced that with 4 cylinders of He gas.
Now that the PPMS is on the system, it's unclear how often we will need to fill. I really don't want to fill
it every week, but it's not looking like we can quite go two weeks between fills. I'm currently filling every
1.5 weeks while I gather more statistics on the boil off and recovery. If I stick with 1.5 week intervals,
that's ~35 fills per year which is probably ~70 L of LHe lost during transfers (with the existing transfer
losses, that's ~100 L per year). I could order one 100 L dewar of liquid He per year + 1 cylinder of He gas;
however, the price with our vendor is essentially the same for gas vs. liquid, so it's probably easier to
simply add one cylinder of He gas to the recovery system every month since we can't get any liquid He from
our vendor at the moment, but we seem to be able to get He gas just fine.
R11: ~ 3600 Liters/year per before installation (Our system is not fully optimized - but we are recovering
-80% of our usage even now)
R12: We used almost 5000 liters before we had the recovery system. We had problems with the system for months
after it was working (partly due to inexperience), so we initially lost a lot of helium. So I only have an
estimate covering the last four months or so. Right now it looks like we're recovering 75-80%. So I think we
still need about 100 liters per month to maintain a steady state.
Q4) Do you liquefy your helium yourself, or do you transport recovered helium to a vendor or other partner for
processing into liquid (i.e., the "hub-and-spoke" model)?
R3: We are collecting helium in bottles and send them to another Uni with liquefier (10 km away).
I am using the service, which they provide (bringing me Helium and I return gas and pay for the helium).
They have a big liquefier and supplying two universities (Chemistry and Physics) and a few research
institutes with their Helium. Here a picture of the equipment:
https://www.physik.fu-berlin.de/service/ttl/index.html<
https://urldefense.com/v3/__https:/www.physik.fu-berlin.de/service/ttl/index.html__;!!BpyFHLRN4TMTrA!4ZlGqW0w1J-Iz_josZT2EfROsZnBo02PXj0AhBg9kgHp-QJYqhbqaRyeOwB2uNNitWUZD1O7R-RqJHvHHIldPTybbfNnnUZ3EdGm$>
They claim to liquify 200.000 l.
R4: liquefy ourselves.
R6: We do everything in house.
When I asked, no vendor I could find would take recovered gas. That could have changed.
TAMU considered campus wide recovery but found it was not worth it to pipe the gas to a central location.
We do have tunnels on campus but putting copper pipes in was really expensive and PVC has issues. I had
not found the DEF tubing at the time. DEF can be bought on spools so it would have been easier to install.
Remote sites would have needed local gas storage and then pumped over, again adding to the cost.
TAMU also considered putting gas in cylinders and transporting to the location. Again cost of installing
compressors at each site and then paying for transport of gas and liquid back to the locations was deemed
too expensive.
R8: We liquefy ourselves. And we are not aware of any vendor in the area who would liquefy He gas.
R9: We liquefy helium ourselves with a 20 L / day liquefier
R10: We liquify the He ourselves.
R11: We use a Quantum Design ATL 160 on site
R12: We liquify the helium ourselves.
Q5)Do you recover the extra helium evolved during magnet fills? If so, please describe special equipment and
practices you use to make this work safely.
R3: No.
R4: yes, we recover during fills. No special equipment needed
R6: In my opinion if you are going to do this you should recover as much as possible and the loss during a
fill is significant. Prepping for recovering fill gas takes a little more planning but not much more in
money so well worth the effort. If you do not capture it then you will still be buying lots of liquid or
gas to make up the loss.
If you are only recovering boiloff then you can get away with much smaller tubing or piping to your recovery
location. Since most NMRs have a KF-25 connection it is very easy to use a 1” tube or pipe for individual
systems. It is recommended to increase this size as systems come together in the pipe system. We used 2"
when the system met up and as the main lines in hallways and vertical chases.
Our piping system travels in drop ceilings and these exhaust lines will get very cold for long distances
during fills. So we would have moisture in the ceilings so we did use heat exchangers to warm the gas
before it went into the ceiling and before the flow meter that we use to monitor the magnet fill process.
This is described in detail with pictures on our website.
These heat exchangers also protect the gas bag from getting cold. And reduce overall moisture on the
components. Thus reducing the risk of corrosion from moisture or galvanic corrosion if you are using
different metals/materials.
R8: Yes the loss of lHe during fills is substantial (30%). We really recommend that this is done. We pretty
much replicated the setup described by Eric Paulson (yale). One modification, which we found important:
Use two way valve at the inlet of the manifold. This allows us to take individual magnets off the recycling
system and vent to the atmosphere. During helium fills we take all magnets off the recycling system. Such
that the fill will not interfere with boiloff of the other magnets.
R9: We recover helium during fills by directing helium to 2” corrugated stainless steel piping before it
enters the recovery system header (2" copper piping). If we want to separate other magnets from the system
during the fill on another magnet, we can open a vent valve on them so that pressure on their magnets is
not determined by the helium recovery system header but by atmosphere. However, after a period of being
very cautious, we generally do not open the vent valves during fills now.
R10: Yes - I followed the design as described on Yale's website. (https://cbic.yale.edu/about-us/helium-recovery )
There are pictures on our website that will make this more clear:
https://chembio.byu.edu/system-details
I have a KF tee on the outlet of the magnet; one side is the normal check valve + tubing to the boil off
flow meter and then into the recovery system through a back pressure regulator; the other side of the tee
connects to a larger diameter flex line that connects to a heat exchanger and then into the recovery system.
When the vent line is closed, the gas is forced to go through the boil off gauge. When the vent line is
opened, the gas go directly into the heat exchanger and then the Omega flow meter and on to the main
collection circuit.
The Omega FL7611 flow meter works great for detecting when the magnet is full (i.e., you can't see "the flame"
since the plume is contained, but the flow rate jumps by ~4 SCFM on the gauge once the magnet is full and
producing a flame). This gauge is also useful for ensuring we don't fill the collection bag faster than
the recovery compressor can keep up (particularly useful when filling a magnet up on the 3rd floor or down
the hall from the NMR facility) - I've found that a flow rate on the Omega flow meter of ~4.0 SCFM (air)
matches the recovery compressor rate (the spec sheet says this is 8.5 SCFM He gas).
R11: We have a 3 way- Tee to divert through a flow meter (Details on Yale Univ website). Magnets at ~ 0.5
PSI above atmosphere and we vent slowly through the flow-meter prior to a fill. Our Absolute pressure
controllers take too long to drop pressure.
R12: Yes, we recover the extra helium evolved during fills. We basically copied the Yale setup. The helium
outflow is plumbed into piping by two paths - a small Tygon type tube for normal boiloff, and large flexible
metal tubing for fills.
***Fills are very scary now! We have popped quench valves on three of our older magnets during fills. (Two
of those are Oxford 500s built in ~1984.) Especially for older magnets, you have to fill at very low pressure,
and be very careful to stop the flow immediately when the magnet gets full. We have a pressure gauge and
flow meter attached to each manifold, and we can monitor the fill using either or both. The flow and the
pressure both shoot up when the magnet is full. We bought a transfer line that separates into two pieces -
one goes into the magnet and one into the dewar. It has a shutoff valve, so we can very quickly shut off
the flow of helium when it's necessary. I would highly recommend one of these for anyone who's going to
fill into piping.
Q6) Please describe your equipment, including the major components and their manufacturers. Was it easy to
configure for NMR systems? Was your vendor knowledgeable about supporting NMR needs? Did you require equipment
not supplied by this vendor?
R3: Balloon: 2 cubic m ; Compressor: Bauer G100-2.2 from 1989; Light barriers and a trigger over a cable
R4: Quantum Design is our vendor. We have the medium pressure system with no bag. The compressor activates
on demand and stores the helium temporarily in 3 large storage tanks. This then runs thru an active purifier
(ATP) with a cold head, and then to the ATL liquefier/dewar. We have two water-cooled helium compressors,
one for the ATP and one for the ATL. We can capture the extra helium from about 2 fills before the storage
tanks are filled and then it takes about a day to liquefy. It all works great.
R6: Equipment: See website for full description including build components.
(
https://nmr.chem.tamu.edu/HeliumRecovery.php)
Each NMR uses a KF-25 connection on the magnet to a standard KF-25 Vacuum hose to the heat exchanger, then
to a flow meter. From the flow meter all gas goes to the recovery room (DEF tubing by Omega Flex) where it
goes through a totalizer so we can monitor SCF of gas coming into the room. Then into a 900 cft gas bag
(Flexliner). A Bauer G120 compressor pulls it out of the gas bag and puts it into 12 HP gas cylinders
(Cramer Decker). Compressed gas is down regulated to 15PSI and goes through a binary gas analyzer so we
can monitor purity of our gas (Stanford Research Systems BGA-244). The gas then goes into 2 Drying towers
and then an LN2 trap (Quantum Designs LN2 Purifications system). This purification system provides UHP
quality gas to the liquifier (ATL-160 Quantum Design). We use a drum scale to weigh our liquid helium
dewars. We have 2 100L Dewars from CryoFab for transporting helium to the magnets.
Most of the vendors are helpful about how to do things but they rely on you to have good numbers about
your systems. Knowing how often you fill, exhaust rates during those fills, duration of fills, total helium
used per system per fill, and boil-off rates. These numbers are crucial for your planning of your plumbing
and how much storage you need and how much helium you need to liquify a day. These numbers may also help
determine if you use high pressure or medium pressure recovery.
Vendors do not supply anything but their products. They will hook their equipment to your plumbing but that
is pretty much it. They will provide all things necessary to connect all their stuff to each other but that
is pretty much it. So in our recovery room they told use they wanted a ½" copper line threaded to ½"
this was for the 15PSI gas coming from the gas cylinders to their purification system. They wanted KF-40
connections off our incoming low pressure lines from the Magnets and also to the line going to the compressor.
These KF-40s need to be about 3' from where the bag was going to be placed. Vendor worked with us on room
layout and bag size to help place things in the correct areas of the room. The 2” line going to the
compressor in the basement reduced to kf-25 again about 4' from where the compressor would sit. They
provided 1 high pressure line of 10' to go from the compressor to our gas cylinder rack. All gas cylinder
connections were our problem and how to get them connect to the ½” line going back to the recovery room.
R8: We have a Cryomech LHeP22 system. Cryomech is quite knowledgeable. Required at a substantial cost was
installation of all plumbing. Additional equipment needed: hoses, dry-scroll vacuum pump, Transfer dewar
for helium.
R9: 6. We have a system supplied by Quantum Technology including a 20 L / day liquefier (QLHe20), a
medium-pressure compressor (HR3), a nitrogen-based helium gas purifier, and medium pressure storage tanks
(equivalent of 60 liquid litres). QT was knowledgeable about NMR systems and informed us about appropriate
vent valves for separating magnets from the system we had to get them ourselves. We also had to construct
a hoist and support for lifting the cold finger out of the nitrogen purifier for regeneration. They told
us the height it had to lift to, but gave no further details, and we were glad we had a mechanical workshop
to build it for us.
R10: We have a CryoMech system with the following components:
Standard collection bag (300 SCF, the equivalent of ~9.4 L of LHe at our altitude).
Medium pressure recovery compressor (8.5 SCFM He gas, air cooled)
8 medium pressure storage tanks (max operating P = 400 psig; at max P this is the equivalent of ~126 L
LHe at our altitude).
Automatic purifier (this has a PT60 coldhead instead of a LN2 bath to freeze out impurities in the He
gas; water cooled)
LHeP22 liquefier (nominal rate is 22 L/day; the factory test was ~24 L/day; my measurements over the past
year show I'm getting ~25-28 L/day; water cooled)
Yes, CryoMech was familiar with NMR systems; they asked for basic information (number of magnets, typical
boil off, etc.) and then recommended a system sized appropriately for the amount of He we generate and our
fill schedule. They specifically design their system (with the atmospheric collection bag) so that it
minimizes back P on the magnets.
Note: CryoMech only provides the liquefaction system. I had to design and build the collection circuit
and the magnet manifolds myself (again, based on the details at Yale's website). CryoMech provides no
support in designing or building the collection circuit and manifolds to connect to the magnets.
R11: Quantum Design ATL 160 with ATP30 purifier - System has 2 Sumitomo F70L Helium compressors that need
to be cooled. We use building water and have a 18 KW Haskris Heat exchanger
You need all pipe and valves etc. up to the purifier and liquifier. They really didn't give a lot of advice
about the piping system. Most of what we learned we got from existing websites and from one-on-one
conversations with people who had been through similar builds. Really there is no one solution. But in
our experience Bruker's recent recommendations on how to configure the system are pretty close on what we
ended up with
R12: Our system is Cryomech. The system works great, but the vendor was not at all helpful with the setup
outside of their system. We required plumbing/flow meters/heat exchangers/etc. that took about one year to
install
Q7) Please describe your recovery line and how your magnets are connected to it. What materials are used? What
size pipe do you use? Does your line include one-way valves, pressure regulators, or other equipment? How do
you check for helium leaks?
R3: Copper pipes 3cm diameter, "normal" back pressure valves on Magnets, additional release valve.
R4: Our magnets are connected to the recovery network with flexible CSST tubing and KF flanges and CSST
adapters. We left in the place the Bruker one-way values directly on the magnet and just connected after it.
We have two pathways, a 3/4" and a 1/2”. We use the 1/2” for boil-off and 3/4” for the fills. They
both connect to a manifold with a large flowmeter that indicates when the magnet is full. In the boil-off
pathway there is a back-pressure regulator, helium flow meter and 1 psi check valve. For fills, the boil-off
pathway is closed with a ball valve. For normal operation, the 3/4” pathway is closed with a ball valve.
The recovery system uses copper pipe.
R6: Again website has this in detail.
We used as much in stainless steel (SS) as we could. DEF (Omegaflex) tubing is corrugated SS covered in a
plastic coating (looks like a standard vacuum line under the coating). This is pretty thick stuff but is
still flexible. DEF can be purchased on spools in whatever length you want. It can be cut on site to length
and has SS compression fittings that have NPT threads for connecting to other fittings.
Came off the magnets with KF-25 (~1” size) vacuum tubing of the appropriate length. This connected to the
heat exchanger assembly via a KF-25-KF-40 reducer. The KF-40 was a butt weld onto 1-¼” copper tubing
since the heat exchangers are 1-¼” tubing base. These were put together by our machine shop to form
a W shape so that the warm gas would be going up into the flow meter and out to the main line. Leaving the
W is again a KF-40 to KF-25 reducer. On top of the KF-25 is a KF-25 cross, one side has a SS pop-off safety
the other goes to a SS one way (0.125 PSI same as old oxford magnets). This one way is connected to a flow
meter to monitor magnet boiloff. The exhaust of this meter is connected back into the system just above
the main large flowmeter on a KF-25 cross via push to connect tubing. On top of this lower cross sits a
ball valve this is only open during magnets fills, this pushes the normal boil-off through the one-way
valve and small flow meter. On top of the ball valve is a larger flow meter (omega same one used at
Yale too) this flow meter is connected by KF-25 to NPT ¾” male fitting of both ends. The flow meter
is topped with another cross that connects in the boil-off on one side and the other has another SS pop-ff
safety. The top of this cross connects to the DEF via a KF-25 vacuum hose and KF-25 to 1”NPT adapter.
All KF was purchased in SS mostly from LDS Vacuum Shopper or Amazon for the tubes.
We do not use back pressure regulators. Between the pop-off safeties, one-way and ball valves we think our
systems are pretty safe from each other. We have 2 rooms with 2 magnets in the same room and our tubing
was in well before our liquifer arrived. So we had plenty of time to test for pressure effects in these
2 rooms before we really got going. We monitored during fills etc. and never saw any signs that the magnet
not being filled was being affected by the other system. Besides these 4 systems that sit next to each
other the rest of our systems are rooms or floors away from each other, so this helps as well.
By using the ball valves before the main flow meters if there is a quench it will force the exhaust out
the magnet quench ports and not into the recovery system. Unless the quench happens during a fill!
We have 2 hand held sniffers to look for leaks (LG Sciences LD-239). The BGA244 analyzer also should let
us know if we have a major leak. We had plenty of time to leak test our plumbing before we hooked it up
to the magnets too. Using a sniffer at the low pressures that will normally be in the recovery lines will
be pretty much impossible. So it is wise to over pressurize to find the leaks while it is not connected to
the systems. We also have valves in various locations in the plumbing so we can isolate areas. This was
very helpful for this testing.
R8: Recovery lines are 2 inch copper. Magnets are connected to it via stainless steel vacuum hoses, which
transport lHe gas during fills. The daily boiloff enters the system via PVC tubing. We follow pretty much
Eric's setup. Our high field magnets use pressure regulators. For the 300- 500 MHz magnet we purchased
back-pressure regulators, but those have not been proven very useful. Boiloff from the magnets is directly
entering the recovery lines. Many operations in the lHe recovery system pressure the recovery lines with
up to 4 psi (for instance when lHe is transferred from the liquefier into the transfer dewar). For these
cases backpressure regulator cannot keep the pressure in the nmr magnet constant. So we take magnets
off-line during critical operation of the recovery system.
R9: The recovery line is 2” copper piping. On the low-field magnets, there is a heavy-duty T between the
magnets and the copper piping. For low-field Oxford and Magnex magnets, during normal boil-off (i.e.
whenever there is no helium fill), helium is directed via the T to the check valve and from there through
the standard Tygon tubing to a plastic three-way valve allowing venting to atmosphere (whenever it is
necessary to separate the magnet from the helium recovery system) or back to the 2” stainless steel
piping leading to the helium recovery system via more Tygon tubing, and also passing through the standard
flowmeter which came with the system. During helium fills, we open the ball valve on the T so that helium
comes out into the other part of the T and straight into the stainless steel piping. For low-field Bruker
magnets, the setup is very similar, but the checkvalve comes before the T and helium always passes through it.
On the pumped magnet, in the absence of a helium fill, the exhaust from the pump is directed into Tygon
tubing to a plastic three-way valve allowing venting to atmosphere if needed, and then into the stainless
steel piping connected to the copper header. Our pumped magnet (Ascend) operates in "heater mode" so
there should not be any exhaust from the APD, except under exceptional circumstances. Therefore we do not
collect from the APD exhaust. There is an extra oil adsorber in the tygon tubing line, with a 50mbar relief
valve so the helium can vent in the event the adsorber becomes clogged. At the magnet exhaust, a T is
attached to the 20mbar checkvalve that is always in place. One side of the T directs flow into the recovery
header during fills (controlled by ball valve). The other side of the T has a 70mbar checkvalve in place
during normal operation. During fills we replace this with a 20 mbar checkvalve.
Could we leave the 70mbar checkvalve in place permanently? The latest recommendations from Bruker seem to
indicate this is fine and are a little different from what we were initially given.
(Initially, as suggested by Bruker, we had replaced the 70mbar checkvalve with a 20mbar checkvalve, and had
not thought through what the repercussions could be. It turned out we had a very unfortunate event when P3
(atmospheric pressure) was ~ 40mbar below the P2- setpoint of 1030. This caused a large amount of helium
to boil out of the magnet over the period of a day or two when the heaters were going flat out to attempt
to keep P2 up to 1030. Luckily no harm done, weather changed just in time, and we had liquid for a partial
fill in the recovery system- but nothing registered on the BMPC2 as being alarm-worthy. EXCELLENT learning
opportunity, but would not want to repeat it!! Bruker had us update the firmware on the BMPC after this.
With the new firmware, the heater power was dialed back a bit, so we wouldn't have lost so much, but we
still wouldn't have gotten an alarm from the BMPC2.)
We check for helium leaks in the header by pressurizing it with helium, then checking to see whether
pressure has been maintained over a period of time. We use a helium sniffer to check for leaks at points
around the magnet.
R10: I used 1" copper pipe (cleaned and capped) for the collection circuit. To avoid contamination (and
the hassle) of welding, we used ProPress fittings. ProPress has a specific o-ring rated for He gas. I
had some concern about the ProPress fittings and He leaks; to test the main collection circuit (e.g., in
case there was a ProPress fitting someone forget to press), I pumped the system down to ~20 mTorr and let
it sit for about a week. There was no detectable drop in pressure.
For the magnet manifolds, I have a number of components (again, based on Yale's design):
-the pipe is mostly 3/4" Cu pipe (again, connected with ProPress fittings and some NPT fittings)
(a) -I have a back pressure regulator (Control Air Inc., 700BP) for the boil off gas
(b) -I have low cracking pressure swinging disc one-way valve (Eclipse 1006A Disc-type Check Valve) -
this is probably redundant since the back P regulator and the normal magnet check valve prevent back flow.
(c) -I have an Omega FL7611 flow meter (for monitoring the plume during a LHe fill)
(d) -I have a heat exchanger (Zephair ZAS040NZA-F) Note - these are NOT big enough. They really
should be the much larger size heat exchangers.
(e) -I have a tee valve for selecting the boil off pathway vs. the fill mode (plume capture) pathway.
This is actually redundant; simply have a normal tee and a shut-off valve on the plume inlet line is
enough (that's how I have the PPMS magnet setup and it works fine).
I still have some concerns about how the ProPress fittings on the vent line will hold up over time due
to the thermal cycling. I have been testing this with a leak detector (a simple Restek 28500) and I
haven't noticed any leaks on the vent lines of the magnets (measured both when not filling and during
a fill while the vent line is frosted over). However, the manifold that I connect to the transfer
dewars when I fill them has developed a leak where I have a brass ProPress fitting connected to Cu pipe.
It's unclear if this was simply a bad ProPress joint to begin with or if the leak developed because
of how often that particular line has cycled through freeze/thaw cycles over the past year.
R11: 1" Corrugated Stainless Steel Tubing (CSST) - brand Tracpipe with Autoflare fittings.
Its simple and fittings are completely re-usable if you make a mistake and have to re-cut pipe.
We have one section of 1/2” Proflex (Home Depot brand) and the fittings were a nightmare to get right.
Every magnet has a 0.15 PSI one way valve located on or very close to the magnet Helium manifold to
protect magnet from back flow in case of downstream leak.
In addition, IT IS ESSENTIAL to have a 70 mBar pressure relief valve in the recovery line upstream of
any pressure regulator or shutoff valve.
Next also ESSENTIAL - in the main recovery line ensure you have 0.25 PSI relief valves that regulate
line pressure anywhere a shut off valve may isolate the line from the recovery system.
You will have to isolate sections of line for pumping down etc at some point, and pressures can rise
very quickly in the main line if they are isolated from the recovery bags etc. (typically 0.5-1 PSI
within a few minutes). With careful planning you may only need one or two. Also get adjustable ones -
preset ones vary dramatically in their cracking pressure and often the line can go above magnet pressure.
We use the McMaster Carr 0-2 PSI regulator (See Yay UNiv website) set at 0.25 PSI.
Have lots of shut off valves so that you can isolate sections of line for testing.
Have additional access points for adding things like relief valves and fitting vacuum pumps.
If you have multiple instruments think about how you will isolate them from each other and the
recovery system when you need to make changes to the system.
Lots of components are often back orders 12-16 weeks but there are things that you can do to make
the system work - just not perfectly.
For pumped magnets - make sure there are relief valves immediately after the exhaust from the pump
(we have ours after the oil filter) but again before any shutoff valve.
Get a restek handled helium leak checker - ~ $1200 - its worth its weight in gold it if you have a
complex system with many connections.
R12: We basically copied the Yale design. It's really great, but overkill. You only need one flow meter,
not two. It's good to make the piping larger (at least 1" in diameter everywhere). It's better to use 2"
diameter flexible metal tubing to connect the magnet to the piping (Yale used 1"). We used their older
design for the flow meter. We modified them by taking out the metal float (which is WAY too heavy), and
put in a painted cork. It really sucks though - half of ours don't work (the cork remains always stuck
at the bottom). I think their updated recommendation (on the Yale website) - an Omega FL-7611 flow meter
that can work without modification is a much better way to go (but I haven't tried it). But again - a
pressure gauge works as well as a flow meter for determining when the magnet is full. Both of them shoot
up simultaneously.
Q8) Please describe where your pieces of equipment are located and whether this was determined in part by access
to special infrastructure like house chilled water or three-phase power. Is it all in a utility room or in the
spectrometer room? Different places?
R3: balloon and compressor in the basement, own small utility room. Ventilation system to avoid low oxygen level.
No water, three-phase power.
R4: Its in the same room as a 500 NMR with cryoprobe. House chilled water and power were needed for both so it
made sense to locate them together. So far, vibrations have not been an issue for the NMR.
R6: All gas from the magnets flows to the top floor (5th) into the recovery room that houses the 900cft gas bag
and LN2 purifier and liquifier. This was chosen because of availability, location and yes power and water.
The HP storage is in the Basement along with the HP Compressor. This was only chosen because of noise from the
compressor. Power had to be run along with all the plumbing.
R8: The recovery system is located in a room next to the NMR lab. Installation of chilled water and 3phase power
was part of a major lab renovation.
R9: All equipment is in the same room as the pumped 800 MHz magnet. This is because the room is very large and
was initially designed to hold a second pumped magnet with cryoprobe. As a result, it had extra chilled water
connections and a good deal of space. The major disadvantage to this setup is the noise of the cryocompressor
in the liquefier. It is significantly louder than a standard cryoprobe. It runs continuously in our setup.
R10: We looked into several options. However, the politics of getting access to the utility rooms was too
painful, so we settled on putting the system in the NMR lab. Luckily, we have tall ceilings, so I could
build a shelf to put the collection bag over the storage tanks & purifier and I had a sample prep room where
I could put the recovery compressor. It's a little tight, but I was able to get the entire system to fit in
a space that's 14 ft x 18 ft.
R11: Same room as NMR magnets in one corner
R12: We have a large room that houses our (pumped) 800, and a 600. We had space inside that large room to
build a room-within-a-room. It's about 16'x12'. All of the compressors are inside that room, and the collection
bag sits on top of it. The transfer dewar and medium-pressure cylinders are right outside the room. Inside the
room we have house chilled water, three-phase power, and lots of normal outlets.
Q9) Please describe your portable dewar(s). Is it equipped with a pressure-building heater? How do you determine
when your dewar is full? Do you have a scale for weighing it?
R3: Pressure on dewars are achieved with He gas bottle or small bladder. Dip stick to measure Helium level
R4: We use the ATL liquefier/dewar directly for fills. We don't transfer to a secondary dewar first. While the
ATL has some magnetic parts, they are not large and it has not been an issue. We have a long transfer line.
R6: We use 2 100L Dewars from Cryofab. They have differential pressure gauges and heaters (we do not use them
but we have them if we need them). The differential pressure gauge is used to ball park the liquid in them
when they are not in use. The gauges are not accurate when doing fills either to a magnet or being filled from
the liquifier.
We do use a drum scale to accurately measure the liquid in them. We use this weight for billing purposes and
calculating liquid used for efficiency.
For filling we usually use the liquifiers level meter to go off what we have used out of it. The differential
pressure gauge on our dewars also has a specific behavior once we get to about 90L full.
DO NOT over fill your transport dewars you will break the vacuum seals!!
So our procedure is to weigh the dewar and then figure out about how much liquid we will need to fill it. We
then hook the dewar up to the recovery system via the dewar dump port with a KF-25 on it. Exhaust will go
through a heat exchanger with a flow meter just like during a magnet fill. We will start the fill and watch
the level meter on the liquifer drop once we have dropped by the volume determined before the fill we stop.
When the exhaust connections have thawed we disconnect them and take the dewars weight again. When the dewar
is going out for a magnet fill it will be weighed before it leaves and when it returns. We keep track of how
much each magnet uses for our records and billing.
Typically we fill dewars to 85 or 90L most of our fills are only 30-50L. This allows for the dewars to sit
for a few days before and after a magnet fill so we do not have to rush to constantly fill dewars.
Dewars are also plugged into the recovery system when just sitting so even their exhaust is recovered.
R8: We have a 100liter cryofab dewar. It is equipped with a pressure-building heater, which did not turn out to
be a good option. Regulating pressure using a Helium gas cylinder is a much more accurate option for magnet
fills. We determine fill levels with a scale; a special ramp was built to easily roll the dewar on the scale.
R9: Our portable dewar is a 200 L dewar (chosen because the largest single fill is 170 L, while the height is
easier for us to manage than a 250 L dewar. We had it modified after purchase to add a pull handle. We do not
have a pressure-building heater or gauge on it (we wish we had a gauge). We have a scale for weighing it.
R10: I have two transfer dewars, both from CryoFab. One is a standard 100 L dewar that uses push gas for
transfers. The other has a heater for pushing the liquid out.
A floor scale is absolutely necessary to monitor the transfer dewar when filling it. Unfortunately, I didn't
think about the need for a floor scale until the basic design was mostly build and I was already over budget;
that means the location of the scale is a little awkward and I could only afford a scale with a resolution of
+/-0.5 lb (which means an error of +/-1.8 L !) Even a 0.2 lb resolution would be better, but I recommend
getting 0.1 lb if possible.
R11: Get a cryogen level meter they are essentially and take the guess work out of cooling the dewar and filling it.
R12: We bought a 150L dewar. It is equipped with a pressure-builder. But for most magnets we don't use it. It
makes the pressure way too high for the older/smaller magnets, so I just use a helium cylinder. I use the
pressure builder when filling the 800.
I don't think the dewar has ever been full. I spend most of my time trying desperately to prevent it from
being completely empty :)
More seriously: we just thump it. We do have a scale, but it's in constant use. And I figure I'd probably
lose as much helium by wheeling it back and forth to the scale as I do thumping it.
OPERATIONS
Q10) Please describe the time and effort associated with operating and maintaining your system. Do you have any
best practices to share? How many people are involved, and are they NMR staff?
R3: My technician is changing the gas bottles regularly (1-2 a week). Maintenance of compressor is done by
a company (annually).
R4: Too soon to know how much maintenance will be required. But the filling schedule is now dictated not
by when the magnet needs a fill, but by when the ATL is full and room must be made for boil-off collection.
R6: When the system is up and running normally we typically do 2 transfers from the liquifier to transport
dewars a week. Lets say 1 hour each.
With DNP and number of NMR's we have we are typically doing 1 magnet fill a week. A few minutes to check
it out and in. I don't count our Facility managed NMRs since we would be doing the fills either way.
We change our LN2 Trap every 4 weeks. About 45 min to change since we slowly drop it into the dewar.
Regeneration of the LN2 Trap is a passive warm up and then 4-6 pump and purges. We have 2 traps so we can
do this when ever and do not stay with it when it is pumping so a few minutes per cycle to swap gas and
vacuum.
We do our own cold head swaps and other repairs/maintenance. Lots of down time with some of this warm up
or cool down over days but you don't sit with it. But actual work time lets say 5 days a year.
We have 3 NMR staff members here but only 2 of us work with helium recovery.
If you are going to use an LN2 purification system. I strongly suggest a monitor of some kind to monitor
your purity. We use the SRS BGA-244. This will help you decided how long your trap will last and when to
pull it. We average about 0.1% impurities coming from the gas cylinder storage bank. Our trap system can
handle 0.2% impurities over 60 days at 10L/min flow. We could change our traps on a larger interval but
the 4 weeks works for us.
R8: Operating the system was a new "hat" for the NMR manager. In overall, the operation of the system will
require some time efforts: lN2 fills of the purifier are required about every 5 days (not very practical
since the fills cannot go into the weekly fill cycle of NMR magnets). The system should be monitored daily,
filters should be changed. Especially changing the filter in the He purifier is some work.
R9: Four people (two NMR managers and two electronics technicians who have always helped with helium and
nitrogen fills, and general NMR maintenance and service) maintain the system. With the help of Quantum
Technology, we have connected a computer to the control interface (HMI) via Ethernet so that we can view
the interface on the computer via TigerVNC. We then use Anydesk to view the computer. We usually check
on the system via this computer twice per day to make sure that pressures are as expected and helium
increase is normal. We end up transferring helium from the liquefier to the portable dewar at least every
two weeks and this requires someone to change one setting in the control interface in order to depressurize
the liquefier a couple hours before the actual transfer, which itself takes up to an hour. We regenerate
the nitrogen purifier every few months (this takes one day, although most of it is passive warmup and
cooldown) and we do maintenance on the compressor and water filter annually.
The twice-daily check on the general system has been found necessary because our liquefier occasionally
shuts off due to perceived problems with the chilled water. QT is supposed to send us a new temperature
sensor for the compressor, which we hope will stop this problem. (A cryoprobe compressor on the same
chilled water system never shuts down when this happens). We have also had other strange events, such
as occasional high boiloff from the transfer dewar when the transfer line is left in it.
R10: I'm the only one operating the system (I'm also the only full time person running the NMR lab - I have a
5 hr/week TA who does LN2 fills). Once I have everything optimized, I should be able to train our liquid
N2 TA to help with some things and the graduate students in the three research labs could do the LHe fills
(if I trust them enough to do it right and not waste a lot of LHe).
It was planning the system, setting up the system, and initially figuring out the best way to do things
that was quite time consuming. Now that I've had the system is up and running for a year, it doesn't
take much time to keep things running:
Every day (~5 min), Check the Vol in the collection dewar and the P in the storage tanks to see if I need
to close the inlet valve. I prefer to keep enough P in the storage tanks that I can push liquid out of
the collection dewar when it's time to fill a transfer dewar. I also prefer to keep the collection dewar
a little below 150 L as I find it easier to push the liquid out if it's not completely full.
Once a week, I record a few things (~30 min)
Measure the boil off rates of all the magnets on the system.
Measure the total He in the recovery system (i.e., weigh the transfer dewars, empty the collection bag and
record the P in the storage tanks and get the Vol in the collection dewar of the LHeP). I compare this
with the avg boil off readings so I can compare the amount of He I'm collecting with the expected weekly
boil off amounts.
Check the color of the beads in the water adsorbers to see if they need to be swapped and regenerated
Record the pressures and total operating hours for the LHeP and the purifier
Periodically (depending on fill schedule)
Top off transfer dewars to keep them cold if there's a long gap between LHe fills (~1 hr)
Top off a transfer dewar the day before a fill (~1 hr)
Top off a transfer dewar after a fill if it's almost empty or will be sitting idle for a while before the
next fill (~1 hr)
Record the total amount of He in the system before and after a LHe transfer to help estimate transfer
losses (~10 min)
The LHe fills take the same time it took before we installed the recovery system; however, I am currently
filling more magnets than I was before (i.e., the other magnets that are not part of the NMR facility)
because I don't currently trust random graduate students to fill things carefully and not waste LHe during
the fill. Once I have everything optimized and we no longer are in the middle of a He shortage, I'll be
more willing to let other people fill their own magnets as it will be less problematic to add He to the
system if someone screws up. This takes 2-4 hours per month depending on the fill schedule.
Every 2 months
Swap the water adsorber on the recovery compressor and regenerate (~1 hr)
Pump down the vacuum space on the cold head in the automatic purifier (~1 hr)
Leak check various parts of the collection circuit and magnet manifolds
Once a year, analyze the data: average boil off rates, average transfer losses, estimated time to replace
oil adsorbers and refurbish cold heads, etc. (~several hours)
R11: We are still working on this
R12: I was basically conned into taking care of the system by myself (take care of the helium recovery system
they said, it won't take much time and it'll be fun they said...). I am NMR staff. The thing seems like it
requires constant maintenance, but realistically it probably only takes ~20 % of my time (which is what Eric
Paulson told me). As far as best practices: almost all the losses in my system seem to be associated with
transfers and fills. It's not unusual for me to transfer 80 liters into the transfer dewar and have 10 liters
or more disappear into thin air. You can measure how much is in the plant, how much is in the cylinders, and
how much is in the transfer dewar. And the total amount of helium is very clearly less after some transfers.
This phenomenon seems to be worse if you do your transfers when the transfer dewar is empty or very low, or
when you don't have much helium to transfer into it. If the transfer dewar has a lot in it (50 liters, say),
and you transfer 100 liters into it, then usually you don't lose any. I don't understand it, but it's been
pretty reproducible for me. So I would say: try to keep as much helium in the system as you can. And of
course, don't let the transfer dewar get completely empty. Or if you have to empty it, make sure you have a
decent amount of helium to put back into it right away.
Q11) If you are comfortable discussing the topic, how was purchase funded? Institutional funds? Government grant?
Other? Did support for different aspects of the project come from different sources?
R4: An NIH supplement paid for the equipment. But the infrastructure was 3x the cost of the equipment and this
was paid by the institution.
R6: NIH supplement funded most of the purchase from Quantum Design.The University took care of most of the
remaining, and small purchases were made from the Facility budget over time as needed.
We saved money by being able to run the DEF tubing pretty much by ourselves. Did hire a company to core holes
in walls etc, and run the tubing in the vertical chases for safety reasons.
R8: Funding came with a major upgrade of the NMR lab (purchase of a 700 and 800 MHz instrument), which was
associated with a major lab renovation. Given all the costs of instrument and building renovation the cost of
the recycling system was not too significant.
R9: The project was funded by a combination of institutional funds (university sustainability funds, Faculty of
Science, and Department of Chemistry) and a government grant for infrastructure.
R10: Several years ago, we hired someone who was bringing their SS-NMR magnet with them. They had a He recovery
system at their previous institution and refused to accept our offer unless we also had a He recovery system.
In the end, part of his startup package paid for part of the recovery system with the college and department
using capital equipment and renovation funds to pay for the remainder of the system + renovation costs. It
definitely helped that our current dept. chair owns the FTICR and knew how painful the past two He shortages
were, so there was a lot of support to get a system in place before another He shortage.
R11: NIGMS supplement for Helium Recovery
R12: We got the NIH administrative supplement back in 2019/2020. That paid for the system. But we were already
in talks with our dean, who was basically ready to pay for everything. The grant was for ~$250,000, which paid
for the Cryomech system, but the cost of the plumbing was about double that. Our dean paid for that part.
Q12) Please describe how you pay for ongoing expenses like maintenance contracts, staff time, consumables, etc.
Did you increase recharge rates to pay for them? Are your costs offset by savings due to reduced helium consumption?
R3: We pay for the maintenance of the compressor, the recovered Helium gas is set off against the delivered
liquid helium
R4: yes, recharge. yes offset by reduced helium consumption.
R6: No contracts since we do it all. Staff is all NMR.
We do charge per liter used by the different groups getting helium from the facility. But we also give a
credit for the helium recovered from each groups systems. We do this by having totalizers plumbed in a
strategic points and taking readings once a month.
Yes our overall cost for helium is greatly reduced by having this system.
R8: Operational costs are quite reasonable (electricity does not count for the budget of an academic lab). So
far the only real supply, which we needed to purchase were ground-pins needed for the sensor measuring the
purity of he gas and a one time change of the adsorber. Cryomech does not offer service contracts. But in the
foreseeable future our system will need a major service requiring rebuild of the coldhead.
R9: We were lucky in our funding applications and were able to pay for the first five years of maintenance
with the initial purchase. After that time, we expect to be able to pay for ongoing maintenance out of user
fees, since we will not be purchasing much helium. Nitrogen, required for our purifier, is also paid for
from user fees. We have not yet reduced user fees because of purchasing less helium, but expect to do so
in the future. Although we have been surprised by how much staff time has been required to monitor and work
with the helium recovery system, we have not increased staff costs.
R10: I have a modest budget that can cover small consumables (compressed He cylinders to cover annual transfer
losses, replacement ground pins for the automatic purifier, o-rings, fittings, etc.)
Eventually, there will be more expensive costs (replace oil adsorbers at 25,000 hours; refurbish cold heads
at 40,000 hours). The department will use capital equipment funds for this.
According to the calculations/projections that I've made, we're saving ~$23k per year (this calculation looks
at how much we would be spending on LHe annually plus the expected cost of replacing the PPMS cold head
averaged over the typical lifetime of the cold head minus the expect maintenance costs of the automatic
purifier and LHeP averaged over the expected time between oil adsorber replacement and cold head
refurbishment). The total capital costs were $309k including the renovation costs. This means we recoup
the capital investments within ~13 years.
Note: if we didn't have the PPMS on the system, the savings is only ~$18k, which is then ~17 years to recoup
the capital costs.
A nice way to look at the savings is to do all these projections (i.e., expected yearly losses + expected
maintenance schedule so those costs can be divided out into a per year cost, etc.) and generate a cost per
L of LHe delivered (i.e., you have to liquify more than you end up delivering due to the vaporization %).
The cost per L will depend on whether the facility has to pay for electricity. For my facility, we end up
with
Just the 5 magnets (4 NMR + 1 FTICR)
$3.81 per L delivered (without electrical costs included)
$8.65 per L delivered (including electrical costs)
All 6 magnets (i.e., with the PPMS magnet)
$2.60 per L delivered (without electrical costs included)
$6.29 per L delivered (including electrical costs)
Compared with what we were paying for LHe in 2019:
$25 per L delivered for a 60 L dewar
$23 per L delivered for a 100 L dewar
$21 per equiv. L of LHe for a cylinder of He gas (this doesn't include the cost of electricity to
liquify that gas)
R11: We don't plan to make changes in charges. What we save in helium costs we offset with additional costs
for cold-head maintenance.
R12: We don't really have a plan to pay for maintenance. We were going to recharge the helium, but we were
disallowed from doing that by our local bureaucracy. Right now we're just splitting the cost of any helium we
do purchase based on instrument usage. We'll have to figure out how to do something like that with maintenance
Q13) Was the cost of infrastructure improvements a major barrier to acquiring a recovery system? Were you able to
use grant money to make infrastructure improvements like installing pipes?
R4: If the institution knew the costs were as high as they were, they likely would not have committed to the
supplement. No grant money was not allowed to be used for infrastructure is my understanding.
R6: To my knowledge our NIH supplement could not be used for infrastructure it had to go to equipment purchase.
We had most of the power and water in place already.
Plumbing was the big obstacle. Cost projections were over 300K to install copper piping for our project. The
DEF tubing and installation costs were about 60K.
R8: For GT the purchase went without Grant money.
R9: We used institutional money for infrastructure work such as installing pipes and installing a 460 V transformer.
R10: The renovation costs were ~$57k. Some of this was absorbed into the startup package of a new hire (the
University has a separate budget for renovation costs related to new faculty hires).
R11: We did a lot of the work ourselves. If we had contracted this to facilities it would have been prohibitive.
Also we found that despite months of planning things were changing all the time and this would have been
difficult to plan for.
R12: It would have been a barrier. Fortunately we have a dean (who's also a professor in our department and a
user of NMR) who was on board with funding a helium recovery system. I don't think we could have used the
grant money for any infrastructure.
Q14) Has your helium vendor treated you differently since you began recycling helium? Have you lost "preferred
customer" status?
R3: We rarely buy Helium from someone else than the neighboring university. But in emergency cases one of the
vendor can deliver helium relatively quick (I am not in the US)
R4: Not yet!
R6: We still order an occasional 60L dewar just to stay on the radar.
I have heard this to be an issue for some. It may help that TAMU as an University still purchases a large
amount of helium?
R8: This is tough to say. Currently our vendor (Airgas South) appears to be happy about every amount of lHe
we are not buying. As the NMR lab we are only one major purchaser of Helium within Georgia Tech, which has
an exclusive contract with Airgas.
R9: No. It may help that our recovery rate was not great (~50%) at the beginning, and we have sold some of our
recycled liquid to other labs when they had trouble getting deliveries. Therefore the amount we are purchasing
has been slowly decreasing since the installation of the recycling system. We only started being able to access
(less expensive) helium from one vendor shortly before installing the system, so they do not have a long history
of our purchasing habits. Our long-term vendor was someone skeptical of how useful a recovery system would
be… and despite us purchasing much less due to be able to buy from the university's preferred vendor, and
having the recovery system, is still supplying us with the same delays as other customers.
R10: I'm not sure as I don't deal with them directly - our purchaser in the chemistry stockroom told me that
they were initially surprised last year when we were no longer ordering LHe and they asked if they had done
something wrong to make us switch suppliers.
In the end, we really didn't have any useful status before - in the previous shortage, AirGas simply refused
to provide us any liquid He dewars at all and we had to purchase our LHe from Praxair for the year or year
and a half that the previous shortage lasted.
R11: No
R12: I don't think they know that we have a recovery system. But right now we're having a hard time because
they're basing our current allocation on our previous one year of purchases, which were pretty low.
ANYTHING ELSE?
Q15) What other things do you wish you had known before getting this system? What mistakes would you caution
your colleagues to avoid? Please provide your best additional advice to people seeking a new helium recovery
system.
R4: its actually all working without any issues. So it has gone better than my most optimistic expectations.
Lucky me!
R6: This was my second Recovery system install so we did a write upon our website about it
https://nmr.chem.tamu.edu/HeliumRecovery.php. I am happy to talk to people and help anyway I can, no reason
to reinvent the wheel here.
PLANNING, PLANNING, PLANNING
Know your systems. How often are you filling, how much liquid each fill, what are your exhaust rates during
your fills, how long are your fills, what are your boil-off rates
Plan for extra storage you will need to shut down for coldhead changes and other issues. Depending on how
you do this it can take 3-5 days so you need to have storage space for that amount of boil-off. You will
also need space during these time periods to hold the exhaust from the transfer you just did out of the
liquifier to get it empty so you can warm it up. That could be 10-40L of liquid worth of gas. Did you already
have some down there? Did a magnet fill need to be done right before you shutdown? Plan for space and plan
the shutdowns as best you can.
Plan for expansion. New systems extra gas storage space.
If using a gas bag can it hold a helium fills worth? Depending on if you are using medium pressure or HP
this will be a factor in size of the bag! Medium pressure pumps are much faster than the HP ones so they
can move the gas out about as fast as you can put it in during a magnet fill. The HP pumps can not!! Ours
is rated for 7.5SCFM of air that means at best 3.5 SCFM of helium. Our slowest rate during a transfer is
7 SCFM so the HP pump will not keep up you can run out of space really quickly. We end up with about
350-400SCF of gas during a transfer. With a 900 SCF bag I can do basically 2 a day without danger. When we
do this we typically empty the bag between them.
The gas bag is typically your largest source of contamination. So we empty our bag routinely when it is
about 50% full to minimize time spent in the bag. We try to once a week manually drain the bag as low as we
are comfortable. Again to help get contaminated gas out so it does not sit.
STRONGLY recommend you get sniffers for leaks and purity monitor of some kind. Really helped us find our
issues in the beginning!
Plan for being able to put gas cylinders on so you can get extra gas/liquid when you need it.
If using your own dewars plug them into the system when sitting. Plan to be able to do this with purchased
dewars too!
We typically try to plan so there is 20L in a dewar after a magnet fill is done. This keeps them good and
cold and we do not have to rush to refill it.
Think about the number of systems you have and can you keep a dewar cold! I have talked to a couple sites
that only have 2 or 3 systems and they go so long between fills and making enough liquid that a transport
dewar goes empty. So think about logistics of this. A plus for Quantum Design's here their liquifier is
moveable, so you can fill directly from it.
We do not pay for power which maybe a considerable amount for some places.
Purification can be via LN2 or a coldhead based system. Each has pros and cons. We did not want the added
cost of the coldhead up front and long term. But I do recommend an extra trap if you go LN2 so you have
less down time.
R8: (1) Talk to people who have a helium recovery system already.
If the system is shared with other non-NMR laboratories work out costs, etc.
We do not have good experience liquefying too much bought helium gas. This leads to the accumulation of
impurities in the liquifier (likely Neon gas).
Make sure that your transfer dewar does not run empty. If it is empty and starts warming up a substantial
amount of lHe will needed just to cool it down again.
You should evaluate individual magnets wrto recovery of regular lHe boiloff and boiloff during fills. Not
too many labs still operate a r2d2 type magnet, but boiloff during fills from these magnets cannot be
recycled (we almost quenched the magnet trying it using a specially designed connector).
R9: We wish we had purchased a transport dewar with handle and gauge at the outset. We have purchased a scale
for the helium transport dewar and we wish that we had planned for a scale for the nitrogen purifier also.
(It is hard to find reasonably-priced scales with ramps and sensitivities to small changes in weight so as
to measure helium during transfers. For the nitrogen purifier, the difficulty is finding one with a ramp and
a small enough footprint to fit within the bounds of the scaffolding for our hoist.) We bought a tall aluminum
stepladder for easier access to the top of the transport dewar after the fact, and we are currently purchasing
a gas cylinder cage and stocking up on helium gas so that we always have helium available for emergencies.
We should have planned for soundproofing because of the loud cryocompressor.
One problem we have encountered since installing the system has been transporting helium to other sites
who need it. In one case, we filled a dewar provided by the other site, but we lost some helium since we
had to cool it. Since then, we have shipped our own dewar to other places, but since our loading dock is
almost permanently under construction, the carriers we have found have refused to or been unable to meet
our dock, and have required us to push our dewar to the truck, damaging the wheel. In one case, our dewar
was returned with the safety valve closed so we have now added signs about proper valve position and we
request users to tie wrap the safety valve open.
Our final issue is the difficulty of measuring recovery rates on a short-term basis. (Over a long period
of time, it's easy to see how much helium was ordered and this can be compared with how much helium would
have been needed without the recovery system.) The gauge on our liquefier dewar doesn't seem to be
perfectly accurate and might be non-linear, the scale our transport dewar sits on isn't very precise
(0.5 lbs), nor very accurate (occasionally two measurements close in time vary by a couple pounds), and
dipping helium doesn't always agree with well with changes in weight, or with estimates of how much helium
has gone into a magnet during a fill, and how much boiloff was captured. On the other hand, we found it
very useful to ask the manufacturer of the helium transport dewar to give us an equation for calculating
helium volume from helium weight. It's not straightforward and depends on the dewar size.
R10: There's a lot of things that would have been helpful to know when I was first planning this:
I needed more space and budget for a good floor scale. The floor scale should have a resolution of
0.1 lb (maybe 0,.2 lb at the most).
It's easier to liquefy the He as it's collected and keep the transfer and collection dewars cold. My
original plan did not include the PPMS and I figured I would simply cool down the collection and transfer
dewars as needed and liquify on demand; so, I purchased more storage dewars than I actually need. (This is
a terrible idea! It's much easier to keep the dewars cold and store the liquid.)
I could not get any information on vaporization (flash) % or typical transfer losses from cooling down a
transfer line for a fill. This is necessary information when doing cost projections and schedule
projections. Average vaporization % is ~25% and the typical transfer loss from cooling down a transfer
line is ~1.2 L of liq He for a fill (cooling the xfer line to a flame and then venting the transfer
dewar + venting the transfer dewar at the end of the fill). It should be possible to reduce this if you
build the collection manifold so that you can collect the gas vented by the dewar when depressurizing.
I wish I would have bought a He leak detector at the beginning (I was borrowing a leak detector, but that
quickly became annoying).
A lot of things had long lead times (that was the end of 2018 and early 2019 before COVID made everything
worse). Even simple things like one-way valves from Circle Valve or Cu pipe and ProPress fittings often
had 6-9 month lead times.
One piece of information that I couldn't find (and I asked a lot of people) when I was first looking at
a recovery system and projecting costs and savings, etc. was how much liquid vaporizes during a fill (e.g.,
how much is lost during a fill if you don't collect the plume? how much liquid do you need to liquify
if some % of liquid is lost to vapor when filling the transfer dewar and again when filling the magnet? etc.)
Over the past year, I've found that you typically vaporize ~25% of the liquid that you push out of a dewar
(it varies from 20%-30% depending on conditions). This is important to keep in mind to ensure you have
liquefied enough LHe for an upcoming fill. For example, if the estimated fill volume for a magnet is 40 L,
you can expect to push ~54 L out of the transfer dewar and that means you'll need to push ~72 L out of the
collection dewar. That is, you need to liquify ~1.8 times the amount of liquid that you plan to collect
in the magnet and you need at least 1.3 times the amount of liquid that you plan to collect in the magnet
in the transfer dewar. As long as you're collecting the plume while filling the transfer dewar and the
magnet, this vaporization % (or flash % as it's often called) just goes back into the system, but it does
take time to liquify, so you need to take that into account so you have enough time to liquify almost
twice the amount LHe than expected based on the refill volume of the magnet.
I'm sure there's more, but those are the things that jump out at me right now.
R11: CSST is awesome. Don't do copper. You can make changes quickly on the fly and bend it around
corners etc. If you make a mistake its a quick fix.
Test every section of piping/connector for leaks on the ground before assembling it will make it much
easier (cap sections of line and pressure test them after flushing with He gas). It will save you many
headaches.
MOST Important advice - Really plan where you have relief valves for protecting magnet and regulating
main line pressure. This is the most important advice.
2ND MOST - buy a leak detector - without it you cannot find all the leaks using soap and water. Do not
believe anyone who says you can.
R12: I think I've covered everything. The main thing for me is that I wish we had made the manifolds
simpler - like I said, with only one flowmeter. And I would definitely recommend using the Omega FL-7611
flow meter that is currently on the Yale site, instead of trying to modify one. I would recommend extreme
caution when filling older magnets while capturing the outflow.
ORIGINAL:
Hi AMMRL Community.
At ENC we discussed surveying the community to see what people are doing to
recycle helium in their facilities, with the aim of putting together a general
"best practices" document. There have been many helpful discussion on AMMRL
concerning specific recycling questions, but with interest escalated due to
the current helium availability crisis, it would be productive to develop a
more comprehensive resource. Let's start with this survey.
If you have a helium recovery/recycling system, please respond and describe
your systems and operations, preferably using the questions below as a guide.
If you have chosen to NOT install such a system, it would be helpful for
you to respond by describing the most significant barriers to adoption in
your circumstances.
Helium recycling is a multifaceted endeavor, including aspects of hardware,
infrastructure, space planning, financing, staffing, and more. It is clear
that no single solution is appropriate for every site, so it will be very
informative to hear how different labs have developed solutions that have
worked in their own circumstances (or, importantly, did NOT work well). It
would be especially interesting to hear from facilities that are working
across significant distances or with other institutions (i.e., the "hub-and-spoke"
recycling model). If there's anything you found important that is not coveredl
by these questions, please share your knowledge at the end of your response.
My personal interest is aimed at gaining insights to help our campus develop
a recycling solution. We haven't had one before, and we found we needed to
learn a lot just to get oriented, and we're still determining whether
installation is feasible. One intent of this project is to develop a quick
orientation guide for other people new to thinking about the issue in practical
terms.
Thanks for your help!
Josh
HELIUM RECYLCING SURVEY
1. If you have a website describing your recovery system, please provide a link.
MECHANICS
2. Please describe your recovery network. How many magnets? Are any of them
pumped? Does your network include non-NMR systems? Does it span different
floors/buildings/blocks?
3. Scale: How much liquid helium did you use annually before installing your
recovery system, and how much now? Also, please estimate the overall percentage
of helium you recover.
4. Do you liquefy your helium yourself, or do you transport recovered helium
to a vendor or other partner for processing into liquid (i.e., the "hub-and-spoke"
model)?
5. Do you recover the extra helium evolved during magnet fills? If so, please
describe special equipment and practices you use to make this work safely.
6. Please describe your equipment, including the major components and their
manufacturers. Was it easy to configure for NMR systems? Was your vendor knowledgeable
about supporting NMR needs? Did you require equipment not supplied by this vendor?
7. Please describe your recovery line and how your magnets are connected
to it. What materials are used? What size pipe do you use? Does your line
include one-way valves, pressure regulators, or other equipment? How do you
check for helium leaks?
8. Please describe where your pieces of equipment are located and whether
this was determined in part by access to special infrastructure like house
chilled water or three-phase power. Is it all in a utility room or in the
spectrometer room? Different places?
9. Please describe your portable dewar(s). Is it equipped with a pressure-building
heater? How do you determine when your dewar is full? Do you have a scale for weighing it?
OPERATIONS
10. Please describe the time and effort associated with operating and maintaining
your system. Do you have any best practices to share? How many people are involved,
and are they NMR staff?
11. If you are comfortable discussing the topic, how was purchase funded?
Institutional funds? Government grant? Other? Did support for different
aspects of the project come from different sources?
12. Please describe how you pay for ongoing expenses like maintenance contracts,
staff time, consumables, etc. Did you increase recharge rates to pay for them?
Are your costs offset by savings due to reduced helium consumption?
13. Was the cost of infrastructure improvements a major barrier to acquiring
a recovery system? Were you able to use grant money to make infrastructure
improvements like installing pipes?
14. Has your helium vendor treated you differently since you began recycling
helium? Have you lost "preferred customer" status?
ANYTHING ELSE?
15. What other things do you wish you had known before getting this system?
What mistakes would you caution your colleagues to avoid? Please provide your
best additional advice to people seeking a new helium recovery system.
Received on Wed Oct 26 2022 - 17:38:11 MST