908-788-5550

Note 87: Design and Development of an Automated Direct Probe for a Mass Spectrometer

Home

Article by: John J. Manura and David J. Manura
AutoProbe Design and Development by: John Manos, John Miller, John Manura, Christopher Baker, Dan Lieske, Roland Roadenbaugh. Software Design by: David J. Manura

Scientific Instrument Services, Inc., 1027 Old York Road, Ringoes, NJ 08551
9/12/00

Introduction:

The direct probe was one of the earliest techniques used to introduce samples into a mass spectrometer. The technique is still popular today because it provides a means to perform rapid sample analysis with minimal or no sample preparation. The electron impact (EI) mass spectrometer mode of operation can be used when the sample to be analyzed is relatively pure or when chromatographic separation of a mixture is not required. Direct probe analysis has also been widely used with Chemical Ionization (CI) mass spectrometry. The simplicity of the CI spectrum permits the analysis of mixtures, whose spectra are difficult to interpret using the EI mode of operation.

Applications of the direct probe technique include the screening of drug and pharmaceutical samples, quality control sample analysis, spot analysis of chemical reaction mixtures to monitor rates of synthesis, and the analysis of compounds which cannot be chromatographed. However the technique has been limited in the past by several disadvantages. The technique was not reproducible or quantitative. The direct probe was normally operated by trained mass spectrometer users, and it did not lend itself for use in an open access environment.

Figure 1 - AutoProbe mounted on the Finnigan Trace™ MS from ThermoQuest

This study describes the design of a new automated direct exposure MS probe (DEP) that addresses the previously mentioned limitations of older manual probes. The DEP probe uses a platinum filament wire for the analysis of samples. Samples to be analyzed are dissolved in a suitable solvent and are then injected onto the DEP filament wire using an automated syringe injector. After injection, the sample solvent is evaporated from the wire using a small current. The sample is introduced through a vacuum isolation valve and into the mass spectrometer (MS). Once inside the MS, a current is applied through the DEP filament to desorb the sample into the MS source for analysis. After analysis, the DEP tip is withdrawn from the MS source, heated to high temperature to clean the DEP filament, and then uninjected through the isolation valve so that it is ready for the next sample injection. The system is completely automated and computer controlled. The software to control the automated probe is fully integrated into the mass spectrometer software to provide for a seamless probe-mass spectrometer operation package. The result of this design was an automated MS direct probe that can analyze samples quantitatively at a rate of 2 to 5 minutes per sample. In addition the technique has proven to be easily used in an Open Access environment. The automated direct probe system developed was named the AutoProbe™.

Development of the Probe and Probe Tip

The development of the automated direct probe began with the design of the direct probe and probe tip. The DEP style of probe tip was selected because it permits samples to be loaded via an automated syringe injection system. A replaceable plug-in style probe tip was designed to plug into the end of the probe shaft. The replaceable probe tip consists of a metal filament base with two electrical connector pins and four small coils of 0.005" diameter platinum wire for the filament. The probe shaft consists of a stainless steel ground shaft through which passes the two electrical leads for the filament.

Figure 2 - DEP Probe with plug in filament


The DEP filament was tested on a Hewlett-Packard MS Engine to determine its normal operating conditions. There is no thermocouple or other temperature sensor to determine the temperature of the filament wire. However the HP Engine has a sight glass over the MS source which permits the DEP filament to be viewed while it is heated. The temperature of the filament is controlled by the current through the filament. The filament current was adjusted by a microprocessor controlled current supply that can regulate the current through the filament between 1 mA to 3000 mA (3.0A). Testing determined that the filament begins to glow red at 900 milliamps, glows bright yellow at 1.2 amps and burns out at 1.6 amps. Three separate current ramps are used for the operation of the DEP probe, each of which requires a different filament current. Normal operating conditions for the probe were selected as follows:

Step Filament Current Range Estimated Temperature
Solvent Removal 5 mA to 20 mA 60 to 100 degrees C
Sample Analysis 10 mA to 1000 mA 100 to 1000 degrees C
Bakeout 500 mA to 1200 mA 500 to 1500 degrees C

The probe was mounted on a servo-controlled linear actuator with a 50 cm stroke length, which provides for probe movement. The servo controller permits the probe to stop at the various set positions required for the analysis, including the sample load position and various positions inside the isolation valve and MS.

Vacuum Isolation Valve

The heart of the AutoProbe is the isolation valve (Fig. 3 and 4). The purpose of the isolation valve is to permit the probe tip, which is at atmospheric pressure, to be injected into the mass spectrometer vacuum system while still maintaining a good vacuum in the mass spectrometer. To accomplish this, the direct probe is first inserted into the first probe seal of the isolation valve. An electrically operated solenoid valve opens to permit the accessory vacuum pump to evacuate the area in the isolation valve between the probe tip and the ball valve. A vacuum gauge with pressure readout measures this vacuum. When the pressure gets down below 200 millitorr, the probe is advanced to the second seal in the isolation valve assembly. Lastly, the pneumatically controlled ball valve opens, which allows the probe to be inserted into the mass spectrometer.

Figure 3 - Probe positioned at the first seal in the isolation valve - Figure 4 - Probe passing through the rotary ball valve

Sample Loader

A CTC PAL Autosampler was selected as the automated syringe injection system to load samples onto the DEP wire. This PAL Autosampler is mounted on top of the probe and aligned so that the syringe would accurately inject samples onto the DEP wire. When injected, the tip of the syringe is within 0.002" to 0.010" from the DEP filament wire. This distance permits the samples to be injected onto the DEP wire without physically touching the fragile filament coil. This distance is also optimal for loading samples onto the DEP wire coil. As a result, the DEP filament is expected to have a long life, physical shape will not vary, and samples can be accurately and reproducibly injected onto the filament wire. This point later proved to be the reason this technique was reproducible and quantitative.

Figure 5 CTC PAL Autosampler on the AutoProbe

AutoProbe Console

The AutoProbe console (Fig. 6) was designed to attach to the MS and to contain the AutoProbe components, including a microprocessor, the direct probe and linear actuator, a filament current controller, the isolation valve and components, the PAL Autosampler power supply, a vacuum gauge, and a number of valves and relays. A Z-World microprocessor is used to control the operation of the AutoProbe. There is PC software to integrate the MS software to the microprocessor in the AutoProbe console.

Figure 6 - Schematic of the AutoProbe Console

AutoProbe System Components

  1. linear Thrust column
  2. Probe Rod
  3. DEP Probe Tip
  4. MS Probe Interface
  5. Isolation Valve Assembly
  6. Electronic Vacuum Gauge Pressure Sensor
  7. Isolation Valve Relay Switches
  8. Filament Current Controller
  9. Relay
  10. CTC Power Supply
  11. AutoProbe Microprocessor Controller
  12. Servo Controller for Linear Thrust Column
  13. Electronic Manifold
  14. Autosampler Syringe
  15. CTC PAL column

AutoProbe Software and PC Control

The AutoProbe System has four levels of software that control its entire operation. SIS developed software for the first three levels. The four levels include:

(1) Embedded Software - This resides on the embedded C-programmable microprocessor controller inside the AutoProbe instrument. It directly controls the various hardware components of the AutoProbe, including the servo controller for the linear actuator, the DEP filament current controller, the isolation and rough-pump valves, the vacuum gauge sensor, and the pneumatic valves for the ball valve.

Figure 7 - ThermoQuest Xcaliburscreen with AutoProbe system setup

(2) PC-based Instrument Control Software - This controls the AutoProbe instrument from the PC. It does this by communicating with the embedded software (1) via a serial communications link. The PC software sends commands (such as sample run requests) to the embedded software and reads back status data such as the filament current, vacuum readings, the probe position, and operating states. C++, wxWindows, and COM technologies were used. CTC developed their own software for controlling the autosampler. In a typical operation, this layer of software is transparent to the user.

(3) Virtual Instrument Modules - Each instrument of the AutoProbe System is associated with a plug-in module called an "Xcalibur virtual instrument (VI)." These virtual instruments allow instrument vendors to integrate their own instrument control software (2) into the ThermoQuest Xcalibur software package (4). Virtual instruments can be thought of as (software) abstractions of instrument hardware. They conform to various standards such as how instruments are configured, methods are set up, and instrument operations are synchronized. Therefore, the entire AutoProbe system is seamlessly integrated into Xcalibur according to the Xcalibur instrument model. SIS developed a VI for the AutoProbe, while CTC and ThermoQuest provide the autosampler and MS VI's respectively.

(4) System Control and Analysis Software -ThermoQuest's Xcalibur software integrates the setup and operation of the virtual instruments (3). It provides a uniform interface for setting up methods and sequences, running samples, collecting data, doing library searches, analyzing results, and printing reports. It also provides an Open Access mode of operation.

As a result of this software integration, the AutoProbe is seamlessly integrated into the Xcalibur software package. The AutoProbe is controlled within the Xcalibur screens and the AutoProbe system status is shown in the Xcalibur system screens.

The AutoProbe System

All the components of the AutoProbe™ were assembled into the AutoProbe Console and attached to a Finnigan Trace™ MS system from ThermoQuest. Cables were attached to the AutoProbe console to integrate all the systems together. A serial communications cable connects the AutoProbe to the PC. A second serial communications cable connects the CTC PAL Autosampler to the PC. An IO cable connects the CTC PAL Autosampler to the AutoProbe for syncronizing the operation of these two systems. A remote start cable connects the AutoProbe to the MS. This cable enables the AutoProbe system to start the MS after the sample has been injected.

The standard sequence of operation of the AutoProbe system is as follows.

1. Samples are prepared in solution in a concentration range between 5 ng/ul and 1000 ng/ul.
2. The operator sets up a sequence of samples, methods, and processing actions in Xcalibur and starts a run.
3. The Xcalibur software signals the AutoProbe, CTC PAL Autosampler, and MS to prepare for an analysis.
4. The DEP probe moves to the load position.
5. Xcalibur signals the CTC PAL Autosampler to begin a sample run by loading a selected sample.
6. Between 0.125 ul and 1.0 ul of sample are injected onto the DEP filament wire.
7. After the sample is injected onto the DEP filament, the solvent is evaporated from the sample in air. (Alternately, the solvent may be evaporated later inside the vacuum or not at all.)
8. The probe moves to the first vacuum seal, and the electric solenoid valve opens to evacuate the probe inlet.
9. After the vacuum reaches 200 millitorr (adjustable), the probe moves to the second vacuum seal, and the pneumatically controlled ball valve opens. (If a leak occurs, the sequence is paused, the AutoProbe resets itself, and an error is recorded on the log window.)
10. The probe moves to its final position against the MS source. The MS is signaled to begin scanning just before this position is reached.
11. The Xcalibur software downloads the DEP filament current values to the AutoProbe controller and the filament begins its temperature ramp. (The AutoProbe can do up to three filament ramps.)
12. The MS stops scanning and collecting data once the filament ramps are complete.
13. The probe moves away from the MS source about 1" and does an optional filament bakeout to clean the DEP wire and prepare for the next sample.
14. After bakeout, the probe moves to the second seal in the isolation valve, the pneumatic ball valve closes, and the electric solenoid closes.
15. The probe returns to its home position and waits for the signal for the next sample.
Sample Analysis

Samples of caffeine (MW = 194) and Cholestane (MW = 372 ) were prepared to conduct a number of studies on the automated DEP technique. Several parameters such as sensitivity, reproducibility, cross contamination, memory effects, and quantitation were studied. At the same time the AutoProbe system was evaluated to determine any operational problems, bugs in the software, life of the DEP filament and the life and reliability of the probe seals and other system component problems. The testing on the system was designed to be exhaustive with more than 2800 samples analyzed over a 2 week time period. During one three day period the AutoProbe was operated continuously and about 1500 samples were analyzed with no interruption.

Caffeine samples were prepared in methanol in concentrations varying from 600 ng/ul down to 1 ng/ul by successive dilutions. In a likewise manner, Cholestane samples were prepared in chloroform in concentrations from 1000 ng/ul down to 1 ng/ul for additional studies. Injection volumes varied from 0.125 ul to 1.0 ul for the various studies.

For these studies, 0.5 ul of each of the prepared samples was injected onto the DEP wire. The AutoProbe DEP filament was heated ballistically for most of the samples to obtain total ion chromatograms in the shortest amount of time and therefore produce the greatest sensitivity of analysis. The DEP filament was heated with 900 mA of current for 30 seconds during analysis and subsequently baked out at 900 mA for 15 seconds in the bakeout cleaning step. The total analysis time for each sample was 3.0 minutes.

Additional studies were conducted by heating the DEP filament with a programmed ramp from 0 to 900 mA at a ramp rate of 50 mA per second to study the effect of a slower temperature ramp on the total ion chromatogram peak shape and sensitivity.

The AutoProbe was attached to the ThermoQuest Trace MS. The mass spectrometer was operated in the EI mode (70 eV) and operated in the total ion scan mode. The mass spectrometer scanned from mass 100 to 400 at a rate of 2 to 3 scans per second.

Sensitivity

Direct probe techniques are known to be less sensitive than GC techniques. This is due to the fact that the resulting peaks in the total ion chromatograph from the probe samples are much wider and therefore less intense than a GC capillary peak.

To determine sensitivity of the AutoProbe DEP technique, a series of dilutions of caffeine in methanol were prepared from 600 ng/ul down to 1 ng/ul. At a concentration of 10 ng injected, the total ion chromatogram exhibits a distinct peak for the caffeine of at least 10 times the background. The mass spectrum of caffeine is clearly detected and a library search identifies it as caffeine. The caffeine 194 peak was clearly detectable in the subsequent analysis down to 2 ng but the caffeine peak in the total ion chromatogram was only about 2 times the background at this concentration. These detection limits were extended 10 to 100 fold when the samples were analyzed in the single ion monitoring mode (SIM).


Figure 8 - Detection of Caffeine at 10 ng level in total ion MS scan.

When the alkaloid drug Ibuprofen was analyzed in a similar manner in a total ion scan, the drug was easily detected down to 0.4 nanogram. The top chart below shows the total ion chromatogram. The lower two charts are the extractions ions for the major ions in ibuprofen at masses 206 and 358.

Figure 8 B - Detection of Ibuprofen at 0.4 ng level in total ion MS scan.

When the same ibuprofen sample was analyzed in the SIM mode for the ions 206 and 358, the sensitivity of the technique was increased to less than 0.1 nanogram as shown below.

Figure 8 C - Detection of Iboprofen at 0.1 ng level in SIM MS scan.

Reproducibility

Historically the direct probe techniques are not known for their reproducibility. This is because it is quite difficult to accurately and reproducibily apply a sample to a DEP probe tip multiple times. In addition the DEP wire is often touched, distorted or the coils shorted out when samples are injected onto the DEP filament wire. Changing the geometry of the DEP filament coil, changes its heating patterns and therefore minimizes reproducibility. The AutoProbe eliminates these problems.

Figure 9 Reproducibility of the DEP Technique

To determine reproducibility of the AutoProbe technique, a sample of Cholestane at a concentration of 125 ng/ul was prepared. Then 0.5 ul of this sample (62.5 ng) was repeatedly injected onto the DEP wire. This was repeated 158 times. The mass spec was scanned over the mass range of 100 to 450 and the total ion signal was collected. The sample was analyzed by selecting the 217 ion of cholestane and integrating the area of this ion in the total ion chromatogram. The peak area at mass 217 for each sample of the 158 samples was plotted in the chart below. The reproducibility of peak areas for all the samples was within 10% of each other. This study proves that the AutoProbe DEP technique is very reproducible. This data also suggested that the automated DEP technique might be quantitative, and this was tested below.

Memory, Background and "Cross Contamination"

One of the major concerns about any MS sample introduction technique are "memory effects" and "cross contamination" of samples. A technique is of little use if "cross contamination" of subsequent samples occurs.

To test for the effect of "memory effects" for the DEP technique, a series of Cholestane samples was prepared in chloroform at a concentration of 500 ng/ul. An injection of 0.5 ul of this sample (total 250 ng of Cholestane) was injected onto the DEP probe tip and analyzed as described previously. This analysis was completed 5 consecutive times within a 15 minute time span (3 minutes per sample) with no blanks between samples. After each sample run, the DEP wire was baked out at 900 mA for 15 seconds. This bakeout step occurs when the DEP filament is about 1.0 inches from the mass spec source. It's purpose is to clean the DEP filament wire to eliminate any residues of the previous sample and prepare the DEP filament for the next sample.

After the fifth Cholestane sample was analyzed, a blank injection of 0.5 ul of clean methanol was injected onto the DEP probe tip. The results are shown in the two total ion chromatograms below. Both plots are at the same intensity scale. The Cholestane sample produced a strong, almost overloaded, total ion chromatogram. The blank sample produced a flat baseline indicating no carryover from the previous samples.



Figure 10 - Background check after running 5 consecutive strong samples.

The second chart below shows the mass spec data at 0.18 minutes for the 250 ng of Cholestane and the blank methanol sample from the data in the total ion chromatograms above. The scale on the blank sample is at 1000 times the scale expansion of the 250 ng of Cholestane sample mass spectrum. No peaks were detected in the blank sample which could be attributed to the Cholestane. This study was repeated several times at concentrations of Cholestane and Caffeine between 500 ng and 1 ng with identical results.


Figure 11 - Background mass spec after running 5 consecutive strong samples

The results above prove that the DEP technique does not suffer from "memory effect" or "cross contamination" problems. The high temperature cleaning step removes all traces of any of the previous samples. It is recommended that the high temperature cleaning step be used in all analyses. The running of a blank sample between analyses is optional and is not normally required.

Quantitation

Direct probe techniques have not been known to be quantitative. The DEP technique has never been shown to be reproducible enough to accomplish this for the reasons stated above. However our studies on reproducibility above show that the AutoProbe DEP technique is very reproducible. As a result the following studies were performed to show that this technique is indeed quantitative.

A series of dilutions of Cholestane were prepared at concentrations from 1000 ng/ul down to 1 ng/ul . Again 0.5 ul of each sample was injected onto the DEP filament and the samples analyzed with the AutoProbe DEP technique as described previously. Each of the dilutions was analyzed 5 times to not only determine if the automated DEP technique was quantitative but to once again show its reproducibility. At the end of each sample analysis the DEP filament was baked out in vacuum at 900 mA for 15 seconds to clean the DEP wire and prepare for the next sample. After each set of 5 samples a blank methanol sample was run. All samples were run in the EI mode and in full MS scan mode.

A plot of the results is shown below. The repeated runs at the same concentration form a tight cluster, again demonstrating the reproducibility and accuracy of the DEP technique. The quantitation plot below is linear over about 2 orders of sensitivity (from 1000 ng down to 10 ng). The linearity of the plot does falloff at low levels of sensitivity (< 10 ng).



Figure 11 - Calibration Curve for the Quantitation of Cholestane

The results proved that the automated DEP techniques is not only reproducible, but it is quantitative. Quantitation with a direct probe to our knowledge has not been achieved in the past. This reproducibility and quantitation capability is due to the automation of the system with the exact same conditions being applied to each sample and to the unchanging geometry of the DEP filament. Since the filament coil is never touched or moved, the sample is injected onto the DEP filament coil in exactly the same point for every sample. This opens up many new applications of the DEP technique for the fast quantitation of analytes.

For additional information of quantitation utilizing the AutoProbe technique, see the next application note on the AutoProbe.

Open Access Operation

One of the unique features of the ThermoQuest Xcalibur™ software is the ability to operate the mass spectrometer in an Open Access Mode. The Open Access System permits inexperienced chemists and technicians to submit samples directly to the mass spectrometer for analysis. The samples are then analyzed automatically and the results reported back to the chemist with little or no interaction of the mass spectrometer operator. The results are increased productivity from the mass spectrometer lab with less technical staff requirements.

The AutoProbe was integrated into this Open Access mode of operation within the Xcalibur™ software system. For a sample to be analyzed, the chemist fills out a simple form on a PC screen including the submitters name and sample name. He then selects a method of analysis from one of several standard methods preestablished by the supervisor. After completing the form, the user is instructed where to place his sample. The samples are then analyzed automatically with no operator intervention. The chemist has the option of having the results e-mailed back to him.



Figure 12 - Open Access Screen in Xcalibur for the submission of samples for AutoProbe analysis

Filament Life

When a DEP probe tip is used in the manual mode of operation, filament life is limited to usually between 10 and 25 analysis. This limited life is not normally due to the burning out of the DEP filament, but to the physical destruction of the filament coil caused by distortion of its shape. It is near impossible to load a sample manually without touching the DEP filament coil with the syringe needle and therefore distorting its shape. With the AutoProbe and the automated syringe injection, the DEP filament coil is never touched by the syringe needle during sample introduction. As a result, the DEP filament coil's geometry does not change with repeated injections.

DEP filament life will be shortened by high current. In our studies, we did not take the DEP filament coil over 1000 mA (1.0 amp). At this current, the filament is glowing red and has an estimated surface temperature of at least 1000 degrees C. This should be sufficient temperature to not only analyze most samples, but is sufficient temperature to thoroughly clean the DEP filament coil. As shown in the studies above, we saw no memory effects from previous samples when the filament was used under these conditions.

In the previous two weeks, we have analyzed over 2800 samples with the same DEP filament coil, and we expect the life of the filament to continue for some time.

Seal Life

The seals for the probe shaft are composed of PTFE. We have found this probe seal design to be superior to other seal designs including the spring loaded probe seal used by several mass spectrometer manufacturers. SIS has been using this style of probe seal for more than 20 years in the design of all our MS probes. Previous testing has determined that the life of these seals should exceed 2000 injections. This is due to the large surface area of the seals in contact with the probe rod. To date we have tested these seals with more than 2500 injections of the probe without significant leakage. Further testing of the AutoProbe will continue to determine the true life of these seals.

When a leak at the probe seals does occur, the AutoProbe isolation valve contains a vacuum gauge between the probe seals and the ball isolation valve to detect any leak. When a leak is detected, the analysis is stopped for this sample as well as any subsequent samples until the source of the leak has been repaired. The probe seals can easily be replaced with a seal removal tool included with the AutoProbe.

Conclusion

A DEP Probe system has been developed that is completely automated and computer controlled. The software to control the automated probe is fully integrated into the mass spectrometer software to provide for a seamless probe-mass spectrometer operation package. The result of this design was an automated MS DEP probe which can analyze samples quantitatively and continuously at a rate of 2 to 5 minutes per sample. In addition the technique has proven to be easily used in an Open Access environment.

The major features of the AutoProbe are as follows:

  • Automated Direct Probe for continuous analysis of samples
  • Uses Direct Exposure Probe (DEP) Technology
  • CTC PAL Autosampler loads liquid samples onto probe tip
  • Can run up to 500 samples per day unattended
  • Samples analyzed at the rate of 2 to 5 minutes per sample
  • Plug in replaceable DEP probe filaments
  • Long life probe tip filaments with expected life > 2000 samples
  • DEP filament heated at a constant current or it can be programmed in up to three programmable current run ramps
  • High temperature bakeout of DEP filament between samples
  • Fully integrated with the ThermoQuest Xcalibur software
  • Can be used in Open Access Mode of Operation
  • Long life probe seals ( > 2000 injections)
  • Probe Seals are easily replaced by user
  • Automated vacuum lock isolation valve
  • Vacuum gauge measures vacuum in isolation valves to eliminate insertion of probe into source if seals are leaking
  • Sequence stop if leaks, broken filament or other problems occur
  • Probe method stored as part of the Xcalibur sample method
  • Can use different methods for subsequent samples in a sequence