John J. Manura
Scientific Instrument Services, Inc. 1027 Old York Road, Ringoes, NJ 08551
(Part I | Part II - Increasing Mass Spec Sensitivity)
This article is the second of a two part series which describes the improvements and changes that we have incorporated into our HP 5971 MSD's in order to improve their sensitivity. In the last newsletter in Part I of this article we described improvements in the mass spectrometer or MSD component of the HP 5971 MSD System to improve its sensitivity (1). Part II of this article describes the improvements in the GC end of the system to achieve additional improvements in the baseline signal-to-noise ratios and therefore improvements in the sensitivity of the mass spectrometer. Although this discussion describes improvements made to the HP 5890 Series II GC and the HP 5971 MSD, the suggested changes and improvements described in this article can be applied to any GC/MS system in order to improve its sensitivity and performance.
In our attempts to improve the GC sensitivity, what we are trying to do is lower the background signal originating from the GC as well as improve the resolution of the GC peaks. Any signal or noise originating from the GC will contribute to baseline signal level and this in turn will restrict the level of analytes that can be detected by the mass spectrometer. There are many components of the GC which can contribute to the background. They can begin with the sample itself, the GC septum, the GC injection port, injection port liners and seals, the GC guard column and all the connecting lines. In addition the GC injection technique, the use of cryo-trapping and the selection of the type and size of the GC capillary column itself will also affect the ultimate sensitivity of the system. This article will describe each of these areas and discuss methods to eliminate or minimize background signal in the GC. In addition we will describe improvements that can be made to the GC system to maximize the signal-to-noise ratio for any analyte.
The following chart lists the major areas from which background signal can originate.
1. GC Injection Port
b. Injection Port Liners
c. Injection Port Liner Seals
d. Gas Transfer Lines
e. Syringe Needle
f. GC Column Inside Injection Port
g. Contaminated GC Injection Port
h. GC Carrier Gas
2. GC Oven
a. GC Guard Column
b. GC Capillary Column
c. GC Cryo-Trap
Most of the contributions to GC background originate from the injection port of the GC. These problems originate both due to the design of the GC injection port as well as the parameters under which it is operated. The background originating from the GC injection port can be minimized by careful selection of the replacement parts such as the septa, liners and seals as well as the proper operating conditions such as operating temperature and septum purge. A typical GC injection port is shown in Figure # 1. This figure will be used to describe the various sources of GC background as well as ways to minimize the GC background.
The GC carrier gas is not preheated in many GC's such as the HP 5890 series, unlike the Varian GC which preheats the carrier gas. This cool gas coming into the HP injection port contributes to the cooling effect at the septum area of the GC injection port and the resulting problems of septum area contamination. Preheating of the carrier gas can help eliminate some of these problems and we do indeed see less memory effects and septum contamination in the Varian GC in comparison to the HP GC. We have considered the manufacture of a column gas preheater for the HP injection port, however the cost of such a device would probably not be cost effective.
It is recommended that when injecting liquid samples into the GC injection port, that the syringe needle be injected fully into the GC injection port. This delivers the liquid sample to the bottom half of the injection port, far away from the GC septum. This minimizes the chance of the liquid sample and analytes from condensing in the septum area of the GC injection port. It also eliminates the possibility that the liquid solvent may wash off materials condensed on the septum from previous injections. However when liquid sample sizes greater than 1 ul are injected into the GC injection port, the rapid volume expansion that occurs in the phase transfer from liquid to gas forces the injected sample analytes into all areas of the GC injection port. Therefore keeping analytes from condensing on the septum is near impossible with conventional injection ports and injection port liners.
In thermal desorption and headspace injection techniques, gases are injected directly into the GC injection port through the GC septum. The gas sample being introduced into the GC injection port is usually quite hot. There is therefore the possibility that if the septum area is cool, the analytes in the sample will condense in the septum area of the injection port. If a syringe is used there is also the possibility that the sample might condense in the syringe needle itself. In one instance in which we were analyzing polynuclear aromatics with the thermal desorption technique, the higher boiling analytes condensed in the thermal desorption syringe needle even though the GC injection port was set to 325C. To overcome this condensation of semi-volatiles we developed a new low dead volume injection port liner to improve the heat transfer to the syringe needle and to the septum area of the injection port. This injection port liner is described below.
The selection of the temperature of the GC injection port can be perplexing. On one hand the higher the injection port temperature, the greater the degree of volatilization of the sample and the minimization of the amounts of analyte condensed in the septum area. However lower injection port temperatures result in less septum bleed, less sample decomposition and less column bleed from the GC column inside the bottom of the GC injection port. We typically set the injection port temperature to between 0 and 25 degrees above the maximum temperature that the GC column will be programmed to inside the GC oven.
The most common source of GC background is the GC septum. Most GC septum are constructed of silicone. The common mass spec background peaks appearing at mass values of 207, 281, 267 and 355 can all originate from the siloxanes in the GC septa. In order to study this further we performed a series of studies to determine the degrees of background contamination that can originate from the GC septum from different manufacturers. We utilized our Short Path Thermal Desorption System in the direct thermal extraction mode. Ten (10.0) milligrams of each of the septum was placed inside the thermal desorption tube and then thermally extracted at 200C to analyze the volatiles present in various GC septa. For this study we analyzed about 10 different septa from various manufacturers. The results of the analysis of 3 typical GC septa are shown in Figure # 2. The worse septa that we discovered were the HP gray septa. The Supelco Thermogreen LB-2 septa were determined to be the best GC septa on the market. They produced the lowest septa bleed. Only three minor peaks were present which corresponded to the siloxanes listed below. The Restek Green septa (not shown) were almost as good as the Supelco Thermogreen septa. (Table I)
The mass spec peaks listed above are commonly seen in normal GC/MS backgrounds. It must be noted that these siloxanes can originate from other sources besides the GC septum. The second most common source is the GC column stationary phases. Non-polar liquid silicone phases such as DB-1 and DB-5 can contribute to these peaks and care must be used not to exceed the upper temperature limits of any GC column. The GC injection port temperature should never exceed the maximum rated temperature of the liquid phase on the capillary column. Also using a deactivated uncoated fused silica guard column will eliminate any siloxanes originating from the capillary column inside the injection port. This will be discussed later.
PTFE coated GC septum are available from many manufacturers. These work quite well, but only for the first injection. After the first injection the problem of silicone background will return. We confirmed this breakdown of the PTFE coated septum in a test study we conducted on headspace vial septum . The results are shown in Figure #3. This study demonstrates the use of PTFE coated Septa on GC headspace vials utilizing the LEAP headspace sampler. The headspace vial was heated to 120 C and 2.0 ml of the headspace gas was injected into the GC injection port via a heated syringe and cryo-focused at the front of the GC column using our GC Cryo-Trap. In the first injection, no siloxane peaks were detected. However after the second and third injections from the same headspace vial, the siloxane peaks increased dramatically in intensity. This occurs due to the fact that as soon as the PTFE surface is pierced, the inside of the headspace vial is exposed to the silicone material in the pierced section of the septum. With subsequent injections, this exposure increases, resulting in increased contamination of the headspace gas by the silicone polymers. The same results can occur in the GC injection port using these PTFE coated GC septum.
Silicone peaks originating from the septum can also result from the incorrect selection of GC syringe needles. The use of 20 point needles will core out plugs of the septa and deposit these small chips of silicone into the injection port liner. These small pieces of silicone are now exposed to the higher temperatures in the bottom of the injection port and will continually bleed the siloxanes into the GC column, thereby raising the GC background. Side port needles are recommended for headspace systems and thermal desorption systems since they minimize the coring of the GC septa. Side port needles can also be used for direct liquid injections but are more difficult to inject through the septum unless predrilled septum are used. Also as mentioned above, longer syringe needles deposit the samples into the center or lower portion of the injection port further away from the septa area. As a result there is less chance of sample condensation in the septa area of the injection port and also less chance of the liquid solvent from the injected sample washing contaminates off the GC septa.
A variety of types and shapes of injection port liners are available from many manufacturers. The selection of proper liner is based on the injection type (split or splitless) and the users preference. Most GC injection port liners contribute to poor heat transfer to the septum area as well as to the sample itself. These injection port liners are typically constructed of glass (a poor thermal conductor) and either have a 2.0 or 4.0 mm inside diameter. The large inside diameters are necessary to permit the rapid sample volume expansion when liquid samples are injected into the GC injection port. The glass or quartz material is critical so as to minimize sample decomposition. Quartz material is normally used for applications requiring a more inert surface. These standard glass or quartz injection port liners are sufficient for most applications.
For thermal desorption and headspace applications, we have designed a new injection port liner (S.I.S. part # SIPL10) to provide better heat transfer to the septum area of the injection port and also to the syringe needle. A comparison of the standard glass injection port liner with our new glass lined stainless steel injection port liner is shown in Figure # 4. The new injection port liner is constructed from glass lined stainless steel tubing (GLT). It is necked down to an inside diameter of 0.75 mm in the top portion which just allows for the passage of a standard 0.63 mm diameter syringe needle. The bottom portion of this injection port liner has an inside diameter of 1.0 mm. These small inside diameter injection port liners can be used for small volume liquid injections (less than 1.0 ul) but were specifically designed for use with thermal desorption and headspace injectors in which a rapid gas volume expansion does not take place.
The metal outer liner of these new injection port liners permits the better transfer of heat to the interior of the liner itself as well as to the top of the injection port area. This minimizes the temperature gradient that was demonstrated in Figure # 1. In addition the tight fit of the injection port liner to the syringe needle provides for better heat transfer to the syringe needle and thereby improves the delivery of higher boilers into the GC injection port in the headspace and thermal desorption delivery techniques. This tight fitting area also minimizes the exposure of the septum area to the sample path flow during injection, thereby minimizing condensation of sample on the septum and septum area during sample injection. The result of using these new injection port liners for our thermal desorption and headspace applications has been the ability to analyze higher molecular weight compounds with less contamination of the GC injection port.
In all cases we do recommend a septum purge. The purpose of the septum purge is to minimize the delivery of septa volatiles and condensation materials into the GC column. When used with the injection port liners described above, the septum purge eliminates most of the silicones originating from the septum. With the HP split/splitless injector with EPC (Electron Pressure Control) this septum purge is factory set to 3.0 ml/min. With other HP models without EPC, the septum purge can be set between 0.5 and 5.0 ml/min.
Many users insert a quartz wool plug inside the injection port liner. This is normally used to prevent septum particles from falling down into the injection port liner and plugging the front of the capillary column. As noted by SGE in the 'Mass Spec Tips' section of a previous issue of this newsletter (4), by adjusting the position of the quartz wool plug in the injection port liner so that the syringe needle tip is wiped during the injection, the GC peak shape can be improved, as well as reproducibility and linearity without adversely affecting boiling point discrimination.
This high temperature flow conditioning of the GC injection port purges all parts of the injection port with gas at the highest temperature possible. The hot gas flushes the injection port area including the septum area and also flushes the septum purge and split vent lines. Areas of the injection port that are not normally cleaned are thoroughly cleaned and purged during this high temperature flow conditioning.
Figure #5 compares the GC backgrounds before and after cleaning the GC injection port as described above. The chromatograms were obtained by heating the GC injection port to 250C, cryo-trapping the volatiles eluted from the injection port on our GC Cryo-Trap for 5 minutes and then analyzing the trapped volatiles by GC/MS. The background noise present in the injection port before cleaning consisted of a wide variety of compounds including septa bleed and injection port contamination from previous samples injected. After cleaning, only three small peaks corresponding to the 3 siloxanes originating from either the GC septa or the silicone liquid column phase were detected.
The type and size of capillary column used can also effect the background levels as well as the peak width and sensitivity of the system. The lower the column phase thickness, the lower the column bleed. Therefore whenever practical, thin liquid phases should be used. However as the phase thickness decreases, so does the dynamic capacity of the column. For most applications we recommend a phase thickness of 0.25 microns. Also as the diameter of the capillary column decreases, the sensitivity of the system increases. This is due to the fact that as the diameter decreases, the band width of the peak decreases. Narrower peak widths result in higher peak heights and therefore greater signal-to noise ratios for the same size sample. However, again as the diameter of the column decreases so does the capacity of the column. The standard capillary columns are either 0.25mm or 0.32mm for most applications.
In order to minimize GC background, all columns must be treated with care. Columns should always be operated at the lowest temperature needed to complete the analysis. There is no need to bake out the GC column at 340C every time when only volatile and semi-volatile compounds are being analyzed. When the columns are first installed in the GC oven, they should be temperature and flow conditioned as per the manufacturers recommendation to prepare them for sample analysis. GC capillary columns should never be heated to high temperature without flow through the column. During this conditioning phase, the column should not be hooked up to the mass spec. This will only contaminate the MS source.
Carrier gases used for the GC/MS system should be of the highest purity possible. Hydrocarbon, water and oxygen traps should be used just prior to the carrier gas entering the GC. The use of clean gas with proper filters can drastically lower the chromatogram background and therefore increase the sensitivity of the system. This is shown in Figure # 7. Oxygen is particularly detrimental to the life of the polar DB-WAX or Carbowax columns but should be minimized or eliminated from entering all columns in order to preserve their life. It is very important to eliminate the exposure of capillary columns to oxygen at high temperatures.
Combining these improvements with the MSD improvements described in the last edition of this newsletter will improve the performance of your MSD or other GC/MS systems.
We welcome comments and additional suggestions from all our readers. I am sure that these last two articles are not all inclusive and I probably missed a few points. So we would encourage you to respond with your ideas. Your input will be published in the Mass Spec Tips section of this newsletter.
(2) Improving Injection-Port Performance in Gas Chromatography, LC/GC, Vol 13, No. 1, January 1995, pp 48-52.
(3) Elimination of ÒMemory PeaksÓ from Thermal Desorption, S.I.S. Technical Bulletin No. 3, January 1994.
(4) Reduction of Peak Tailing, Mass Spec Tips, ÒThe Mass Spec SourceÓ, Vol XVIII, No. 4, July 1994, pp 10-11.
(Part I | Part II - Increasing Mass Spec Sensitivity)
Yttria coated filament at start
Yttria coated filament after 16,000 cycles