Dycor Mass Spectrometer

Updated 25 October 2021

Updates: Ion Pump Revival

Contents

Use the browser "back control" to return here.

• Introduction
  
• Features
• Massport
• Operation, Inlets, Mass Spectrum, Multiple Ion Plot, Partial Pressure, Dycor Quadrupole Mass Filter
• Ion Pumped Mass Spectrometer, Power Consumption
• Membrane Inlet Membrane Performance, Revised Membrane Inlet, Ion Pump Revival
• Mass Spectra, Application

Introduction

The Dycor mass spectrometer is a small instrument intended for gas and vapour identification and analysis. The instrument is linear and is sensitive to gaseous chemicals having concentrations from sub ppm to 100% levels. The instrument is portable and has travelled off site by car to solve industrial and environmental problems. A small generator or a battery-inverter combination can be employed for remote use.

Features

• A Molecular Drag Turbo-molecular pumped vacuum system produces a residual vacuum of 2x10^-8 torr.
• The mass range is from 1 to 200 daltons (mass to charge ratio or m/z).
• The mass spectrometer is controlled with MASSPORT software. The zip-file also includes other software useful in an analytical lab.
• With computer control up to 200 masses can be recorded as ion plots along with 0 to 24 analog channels from other instruments.
• The software has a small library of common gases and vapours.
• To control the Dycor Mass Spectrometer the best current computer is a Raspberry Pi. This requires only a minimal setup and is not expensive. MASSPORT is run using DOSBox. The current version is MASPORT5. A USB to RS232 adaptor makes the serial connection.
• A 50 micron fused silica capillary inlet is used for direct admission of gas samples at atmospheric pressure.
• A Silicone rubber membrane inlet is used mostly for process gas analysis or gases with a significant dust or water content. Recently a Peltier cooler/heater has been added so the membrane can be used to trap headspace vapours at low temperatures and release them at high temperatures. The temperature range is from 0°C to 90°C.
• A small vacuum furnace is used for direct release of gases from solid samples into the ion source. This method is extremely sensitive.
• A tubular silicone rubber membrane is used for the direct measurement of gases in water.

MASSPORT

MASSPORT is a DOS based program. It is not out of date, as it can now be run on modern PCs by using the program DOSBox. Pressing "Print Screen" with a suitable imaging program running will cause the plot window to be saved for immediate viewing and printing. The MASSPORT data is also saved in the normal way. There may be some problems running this software on modern 64 bit PCs.

The Dycor system 2000 software requires the older control units to be updated before communication can be established. Presumably this is a PROM upgrade.

I have rewritten MASSPORT so it functions with some later PCs, but not the most recent. The zipped file MASPORT5 also runs on on the Raspberry Pi 3B using DOSBox. This file is now mass spectrometer only, without any additional data logging capability which would require two serial ports. It is probably simpler to log temperature or any other associated variables on another Raspberry Pi computer. Making the temperature change linear and repeatable, with a suitable controller, means that it is often not necessary to log it at all.

Use the monitor command sudo apt-get install dosbox to install DOSBox. I used a USB to RS232 cable and set the DOSBox configuration file to look for TTYUSB0. Enter serial1=directserial realport:/TTYUSB0.

Everything seems to work. My Raspberry Pi configuration file is dosbox-0.74.conf. This is a file that is placed in a hidden folder as: home/pi/.dosbox/dosbox-0.74.conf after first running DosBox. You need to enable the viewing of hidden files using a suitable text editor. Rename the original file to dosbox-0.74-old.conf. Save the modified dosbox-0.74.conf file into this directory.

My latest version of DosBox has a default configuration file named dosbox-0.74-2.conf. In this case the new configuration file should be renamed to dosbox-0.74-2.conf and the original file renamed to dosbox-0.74-2-old.conf

The zip file contains m.bat and Masport5.exe. Create the directory /dos/dycor/exe under /pi. Place these files in the /dos/dycor/exe folder. Run DOSBox. Masport5.exe should now run.

m.bat is a text file which simply states masport5. It can be modified to run other DOS programs by changing the program path and title. The screen resolution of the Raspberry Pi can be reduced so MASSPORT displays larger. The Raspberry Pi runs well for normal mass spectrometer data acquisition. The antilog function now works properly. I substituted the double-precision Turbo Basic statement EXP10(x) for 10^(x).

For me the Raspberry Pi was inexpensive and the monitor was free. I used a HDMI to VGA adaptor to suit the monitor, although many recent monitors have HDMI ports. There is no reason to use a more expensive or obsolete computer in this role.

My MacBook Air also runs MASSPORT without problems using DOSBox. The USB to serial cable driver needs to be installed and the configuration file needs the serial entry serial1=directserial realport:/cu.usbserial-FTBUQ219 Many other systems should work since DOSBox runs on many computer platforms. I am confident that Linux should work. ChromeOS may work.

The text files produced by MASSPORT are named either *.bar for SPECTRET bar-graph spectra or *.dat for TABDVM multiple ion plots. A typical ion plot file might be named Run1.dat for example. DOS only allows up to 8 letter file names followed by up to 3 letters as a suffix. This is an 8.3 file name. I normally record current experimental data in the default /dos/dycor/exe folder. Periodically I transfer the data files to other folders or to a data key. These files can be renamed to *.txt and viewed with a text editor or imported into a spreadsheet.

Operation

A mass spectrometer sorts out gas mixtures into its constituents as well as producing fragments of each constituent of the gas mixture. The gas mixture enters an ion source where charged molecular fragments are produced by collision with electrons. The ions produced are extracted and are sorted out in a mass filter according to their mass to charge ratio (m/z). The ions are detected as electrical signals with an electron multiplier or a Faraday plate, which is simply a hollow metal tube. The ions land on the surface and are neutralised by extracting electrons from the Faraday plate. The current produced is measured with an electrometer or high impedance amplifier. The electron multiplier improves the signal to noise ratio at least 100 times. The electron multiplier voltage is usually set to produce a gain of about 1000.

Inlets

The inlet of the Dycor mass spectrometer is a 1 metre length of imide coated fused silica capillary with a 50 micron bore. This can be effectively blocked if condensed water enters the capillary. If that happens turn the valve off. Remove the capillary along with nut and the Vespal ferrule. Heat in an oven overnight at 150°C and miraculously the capillary will be unblocked in the morning. If two capillaries plus ferrules are prepared then little time is wasted becoming operational again.

The membrane inlet on this mass spectrometer is very simple. A stainless steel frit is fitted into a short length of 1/4 inch stainless steel tubing. A groove is machined near the frit end of the tube. The tube end is smoothed off and polished. A silicone membrane can be fitted by pushing it on with an O-ring. The groove helps to retain the O-ring. If the surrounding space is constrained the membrane can be tied off, with linen thread or dental floss, just behind the O-ring. The O-ring can now be removed. Finish tying off so the groove helps to retain the membrane. A little PVA can be used to further secure the thread. A Swagelok T fitting is used to suit flowing gas streams. It may need to be drilled out, to an internal diameter of no more than 8.5 mm, to accommodate the membrane.

The best solution is to use the Swagelok T fitting to support and fit the membrane in one operation. This could also be achieved by simply drilling out the Swagelok T to about 7 mm internal diameter and pressing the 0.5 mm silicone membrane on directly. The membrane would compress by about 7 - (6.35 + 0.5 + 0.5) = -0.35 mm. At any particular point the compression would only be -0.35/2 mm, or about -0.18 mm. Initial stepped drilling to 7.5 mm diameter would reduce installation forces on the membrane. With a properly machined fitting the membrane replacement would be quick and easy.

Mass Spectrum

Low mass ions are displayed as a vertical line at the left end of a scale while heavy ions are displayed towards the right. The length of a line represents the quantity of that ion. It relates directly to the proportion of that ion forming component in the gas mixture. The horizontal axis is the mass to charge ratio (m/z) in daltons.

It is useful to have a spectral library for reference. Recorded spectra can be compared to this library using computer based search methods. An example is shown above and at right.

Multiple Ion Plot

Another way of recording information is to select particular m/z ratios as being representative of a particular compound. Their variations with time are recorded and in this way several components of a mixture can be followed in real-time. These are called multiple ion plots. A logarithmic y axis shows related ions well.

Partial Pressure

The term partial pressure is often used with this instrument. It is expressed in torr which is 1/760 of a standard atmosphere. It is simply a number approximately related to the real pressure of a specific component in the vacuum system of the mass spectrometer. Sensitivity factors can be derived for each species by calibration and applied to future work.

Dycor Quadrupole Mass Filter

The Dycor quadrupole, shown at right, behaves as a mass filter. The filter has four 6.5mm diameter stainless steel rods with the opposite rods electrically connected. The rods have an ion source mounted above and two ion detectors sited below, namely a Faraday plate and an electron multiplier. Ions created by the ion source are accelerated along the central axis of the rods by a -70 volt electric field.

An oscillating RF/DC electric field is applied to the rod pairs which causes a spiral ion motion around the axis The applied RF frequency is about 6.7MHz with an amplitude up to a few hundred volts. The RF peak voltage is several times that of the DC voltage so an electric field reversal takes place at twice the RF frequency i.e 13.4MHz.

The ion motion may or may not be stable in the electric field. If the ion motion is stable then the ion reaches the detector and it is measured as a small electric pulse. With many ions the pulses add to produce a continuous signal. If the ion motion is unstable, the ion is neutralised on a surface and it is not measured. The rods with the positive DC voltage applied transmit ions above a particular mass to charge ratio. The rods with the negative DC voltage applied transmit ions below a particular mass to charge ratio. The value of the DC or RF voltage determines which ion is selected. The ratio of the RF to DC voltage determines the resolution by slightly overlapping the low mass range with the high mass range.

Ion Pumped Mass Spectrometer

Recently my MA200 mass spectrometer has been off-site, monitoring the output of a pilot plant. I have looked at the old parts I have from defunct equipment formerly used by others. With some component exchanges there was enough to build up an ion-pumped Mass Spectrometer. Previously I built several mass spectrometers in the years from 1988 to 1996. The old components I now have are therefore about 22 to 30 years old. The following notes describe the development of this mass spectrometer. As changes are made I will update the notes.

The vacuum pump is a Varian 911-5005 8 litre per second diode ion pump fitted with a Varian 911-0030 magnet. The power supply is a small +3500 volt Varian 921-0015 controller. These items were purchased many years ago as refurbished items from Duniway Stockroom. The ion pump is connected in place of the normal turbo pump. A backing pump is only needed for start-up. Otherwise the set-up is conventional, apart from the inlet system. A manual for the ion pump is here. Some further information about ion pumps is in the manuals available from Duniway Stockroom. A wide range of vacuum equipment is available from this source. Look under Documents - Data Sheets for Ion Pump operation, applications and troubleshooting guides. Agilent, who now own Varian, also have informative documentation about ion pumps.

Ideally a turbo pump is required for starting an ion pump. Alternatively a Sorption Pump could be used with liquid nitrogen. My method, using a two stage backing pump, is to heat the mass spectrometer while rough pumping. The high vacuum valve is closed and the pump is disconnected. As the mass spectrometer cools an additional vacuum is created which is sufficient to start the ion pump. The clean mass spectrum shows that no significant oil back-streaming occurs.

There may be a delay before the ion pump starts to work. There will be a slow current drop and a corresponding voltage rise which will accelerate as the pump starts to work. If the current rises there is probably a leak present. There may be pauses in the pump-down while adsorbed gases are dealt with. In my case, after 25 years, there was one pause and then a steady progression to residual gas levels below 10^-7 torr.

The high vacuum valve and the backing pump fittings can be replaced with a cold welded copper pinch-off fitting once the high vacuum performance is established. Different backing pump fittings will be needed for the initial evacuation. To pinch off the copper tubing try the tool used by refrigeration engineers for cold welded copper pipe seals. The finished seal is sharp and fragile so it should be covered with heat-shrink tubing.

Power Consumption

The 3500 volt Varian 921-0015 power supply draws 5 watts from the mains at standby pressures. This is much less than a neighbouring LCD display, which draws 23 watts. At 3 x 10^-6 torr the power drawn from the mains rises to about 8.5 watts

The Dycor console draws about 125 watts from the mains with the CRT display on and about 100 watts with the display off. The Raspberry Pi computer and power supply uses 4 watts running MASSPORT. The maximum total power consumption for a working mass spectrometer and computing system is about 160 watts with the CRT display on or 140 watts with it off. This reduces to 5 watts under standby conditions.

In the distant past I used ion pumped mass spectrometers as field instruments in geothermal locations as well as at several industrial sites. The low standby power consumption makes it easy to relocate this mass spectrometer by car while maintaining a good vacuum.

Membrane Inlet

Any inlet system needs to be matched to the ion pump so that the gas admission rate matches the pumping speed. This is more critical for ion pumps with lower pumping speeds. In my case I have developed a simple membrane inlet which allows for standby operation when not in use.

This design uses a KF16 vacuum centering ring with a built-in sintered metal filter. A disk of 0.5 mm thick silicone rubber, slightly larger than the filter, is wedged in place by a 9.8 mm OD washer. The internal diameter is a nominal 10 mm so the silicone rubber is well compressed by the washer. Excess silicone rubber is cut away and some RTV silicone sealant is applied around the edges of the washer to complete the seal.

The active diameter of the inlet, represented by the hole in the washer, is 3.75 mm. The optimum diameter is probably a bit less than this or a second membrane layer can be used. The required exposed diameter will depend on the membrane thickness used. Additional 3.75 mm diameter silicone discs can be cut and added to the membrane in the centre, until the pressure settles to an optimum value. This silicone rubber adheres to itself and is sold in electronic shops as weather proofing electrical tape. Before use, I heated the membrane at 40°C for a few hours to remove any volatile organic compounds.

On the other side of the KF16 centering ring a drilled aluminium plug was machined and pressed in place and sealed with 5-minute Epoxy adhesive. The short 1.5 mm OD stainless-steel line to the mass spectrometer was pressed in place and sealed with Loctite 243. My approach to the use of adhesives in vacuum systems is to use only tiny amounts. Choose materials which don't smell of any solvents. I simply did a little bit of lathe work and used components that were available here.

Membrane Performance

The working pressure is about 3 x 10^-6 torr for two 0.5 mm thick silicone rubber membrane layers with an exposed diameter of 3.75 mm. At this pressure the pumping speed is almost constant over a moderate pressure range so pump characteristics should not influence experiments. Here I am using the mass spectrometer calibration for pressure. The ion pump current suggests the pressure is about 3 times higher. The ion pump is essentially an uncalibrated Penning gauge with an unknown leakage current. When the Dycor mass spectrometers are shipped they are calibrated against a reference Bayard Alpert gauge. The open ion source of the Dycor mass spectrometer is very similar to this gauge.

Running with the membrane inlet continuously in an open state will shorten the life of the ion pump. I devised a simple plug with an O-ring seal which reduces the mass spectrometer residual pressure to well below 10^-7 torr. The rest of the plug occupies the space above the membrane. A short 3.75 mm diameter spigot is machined on the end to lightly press on the membrane surface. The trick is to have the spigot touching the membrane just before the O ring completes it's seal. A little silicone grease on the O-ring helps with the initial sealing. Thereafter the increasing vacuum holds everything in place. There may be an initial pressure pulse when the plug is fitted, which momentarily increases the ion pump current.

There is no need for perfection here. Any reading in the low 10^-7 torr range would do. The aim is simply to move into a pressure range where the pump lifetime is reasonable. The 8 l/s ion pump specification gives an approximate lifetime of 40000 hours at 1 x 10^-6 torr. A low standby pressure also means that moderate power cuts can be tolerated.

When the plug is removed there may be a period of adjustment as the membrane rebounds from the pressure of the spigot. A total pressure of about 3 x 10^-6 torr is attained. There may also be some more argon instability as the ion pump works harder. There seems to be a minimal effect on the other components of air such as nitrogen and oxygen. Other ion pumps such as the Nobel Diode type where one electrode is tantalum, Triode ion pumps or the rebuilt Varian style 8 l/s Duniway Galaxy Diode ion pump may be more suitable.

I made up another plug with a 50 micron capillary fitted. This is in series with the membrane inlet. It works, but it is a bit sluggish in operation. This will depend on the dead volume below the capillary.

Revised Membrane Inlet

A 8.5 mm diameter and 9.6 mm long aluminium cylinder had a 1.5 mm diameter axial hole drilled from one end and a short 5.5 mm diameter hole drilled from the other. A short length of the 1.5 mm diameter hole was enlarged from the base with a 1/16 inch drill. The 5.5 mm hole was made flat-bottomed with a second reground 5.5 mm drill. Two 5 mm diameter disks of filter paper were fitted in bottom of the 5.5 mm diameter hole. A disk cut from a thin aquarium air filter mat would also be suitable. An oversize 7 mm diameter membrane was pushed into the hole by a 5 mm diameter washer (used for 2 mm diameter bolts) and a 4 mm diameter flat bottom punch. These items were arranged concentrically beforehand.

Surprisingly, the membrane filter assembled itself with an even membrane rim protruding from around the edge of the washer. The edge clearance meant that the 0.5 mm thick silicone membrane was compressed by 50 percent near the edge. Washers are usually smooth on one side and sharper on the other, from punching them out from sheet metal. I used the smooth side to press on the membrane. The compressed membrane edge helped to provided a good gas-tight seal. A 1/16 inch diameter tube was pressed into the other end, using a vice, and sealed with loctite 243. A stainless steel nut and ferrules completed the inlet. This membrane inlet is essentially a smaller and simpler version of the original.

A machined cap, complete with entrance and exit ports, could be easily fitted to allow flowing sample gas streams to be sampled. Gas entry and exit would be on opposite sides just above the top edge of the existing membrane inlet fitting.

Membrane inlet gas admittance is higher at increased temperatures. In previous designs I used a small thermostatted and insulated peltier-cooled temperature controller attached to the membrane inlet. For many routine applications the stability of the working environment is sufficient.

To close this membrane inlet, a 2 mm diameter rod, with a smooth polished end, could press on the membrane surface to reduce the gas transmission rate. Some force would be needed, like any other valve. A stepped acetal plug, with an O-ring seal, could be machined which would consume most of the volume above the membrane. This would replace the simple silicone rubber septum plug shown above.

I am currently using a plug which is a rubber pad normally used to cover the screw-heads on small plastic electronics boxes. The base pressure is about 2 x 10^-7 torr as measured by the ion pump power supply. Note that in small vacuum systems there is unlikely to be total agreement between some pressure measuring devices, as they also act as local vacuum pumps. The mass spectrometer usually displays lower pressures than the ion pump power supply.

I have tested this setup to see if the system could be shut down for an extended period and then successfully restarted, without the help of a backing pump. After a 24 hour shutdown the ion pump was back at its normal base pressure in 15 minutes. Longer shutdowns are possible, but the time taken to attain the base working pressure again will also be much longer. It all depends on the degree of residual leakage from the membrane inlet with the plug installed.

Following a recent 5 day trip, I was able to restart the ion pump when I returned. I turned the power supply on and waited 30 minutes to let the vacuum stabilise. This performance depends on the scale of any residual leaks. I plan to use a small hydrogen fuel cell to look for any remaining leaks using the mass spectrometer set to follow m/z 1 and m/z 2.

Ion Pump Revival

After two years of continuous operation the base ion-pump current rose from below 10𝜇A to just over 20𝜇A. A prolonged bake-out at 125°C liberated gases implanted near the ion-pump electrode surfaces for eventual deposition at greater depth. The ion pump current was monitored at 125°C. A current drop was observed which eventually settled at a steady value. The residual gases were slowly being embedded in the electrode surfaces at a greater depth than usual.

After heating for a day at 125°C the base pressure, at a room temperature of about 19°C, dropped to 8 x 10^-8 torr. This was equivalent to an ion-pump current of 8𝜇A. The following day I repeated the bake-out with no change in the base pressure observed.

Eventually the ion-pump will lose capacity. The pumping speed will be reduced and some instability may be observed. The ion-pump will then need to be rebuilt or replaced.

Mass Spectra

The residual mass spectrum of an ion pumped mass spectrometer can look a little unusual. The spectrum has masses for hydrogen (2), water (2, 16, 17, 18), Air (28, 32), argon (40) and carbon dioxide (44). There are traces of nitric oxide (30, 14), carbon monoxide (28, 12), residual hydrocarbons and secondary peaks for the above gases. Note that the residual pressure is very low at about 6 x 10^-8 torr. The residual pressure can be reduced further after warming the mass spectrometer to about 100°C for a few hours. The following mass spectrum showing residual ions was emailed from my Raspberry Pi computer.

Species such as carbon monoxide, nitric oxide and some carbon dioxide are created by the Penning discharge in the ion pump. There are carbon inclusions in the titanium plates of the pump as well as residual hydrocarbons and water. Argon is elevated because there is no chemical reaction with titanium in the ion pump. Burial of argon by metal sputtering is the only pumping mechanism. There may be some argon signal instability as already buried argon is sometimes released due to sputtering of titanium metal. Triode pumps are more stable in this respect as the pump envelope can act as a surface to trap neutral gases with no sputtering. In the residual spectrum carbon dioxide is elevated because of water-carbon reactions in the Penning discharge of the ion pump. At working pressures the membrane inlet admits carbon dioxide at about 5 times the rate of many other gases, probably because of the linear molecule shape.

Application

For evolved gas experiments from heated samples I would use an aquarium pump, a silica gel drier and a small temperature controlled tube furnace. The exit flow would be cooled and passed over the membrane inlet. The transparent cover shown can be machined and drilled to suit. When not required the ion pump can be put into a standby state by inserting the plug.

It remains to be seen if a small 8 l/sec ion pump can perform well, long term. Routine operation requires some care in use. Reading the ion pump manual is very helpful. My old 20 l/sec Varian triode ion pump was very useful in the past, but I do not have a suitable power supply. So far the ion pumped mass spectrometer has worked well, and silently, with a much improved inlet.