Int. J. Ion Mobil. Spec. (2009) 12:131–137 DOI 10.1007/s12127-009-0028-7
Comparative measurements of toxic industrial compounds with a differential mobility spectrometer and a time of flight ion mobility spectrometer Bert Ungethüm & Andreas Walte & Wolf Münchmeyer & Gerhard Matz
Received: 5 May 2009 / Accepted: 8 September 2009 / Published online: 25 September 2009 # Springer-Verlag 2009
Abstract Small concentrations of toxic compounds in atmospheric air have often to be measured selectively by portable equipment. Ion mobility spectrometers are instruments used to monitor explosives, drugs and chemical warfare agents. First responders also need to detect hazardous gases released in accidents while transporting them or in their production in chemical plants. Not all toxic gases can be measured with the time of flight ion mobility spectrometer at concentrations required by safety standards applied in workplace areas. The time of flight ion mobility spectrometer is based on an inlet membrane, an ionization region, a shutter grid and the drift region with a detector in the drift tube. The separation of ions is due to the different mobility of the ions when they are exposed to a weak electric field (E=200… 300 V/cm). High field asymmetric waveform spectrometry or differential mobility spectrometry is a relative new ion mobility spectrometer technology. The separation is due to the different mobilities of the ions in the high (E = 15000...30000 V/cm) and the weak electric fields. About 30 different toxic industrial chemical compounds were analyzed with both systems under comparable conditions. For selected examples the detection limits, the selectivity and the identification capabilities of the two systems for some of the main compounds will be discussed. B. Ungethüm (*) : A. Walte : W. Münchmeyer Airsense Analytics GmbH, Hagenower Str. 73, 19061, Schwerin, Germany e-mail:
[email protected] G. Matz Institute Of Measurement Technology, Hamburg University of Technology, Harburger Schlossstr. 20, Hamburg 21079, Germany
Keywords Ion mobility spectrometry . Time of flight . Differential mobility . Toxic industrial compounds
Introduction Ion mobility spectrometers (IMS) are widely used for monitoring an extensive range of hazardous compounds [1, 2]. Applications in the military e.g. the detection of chemical warfare agents (CWA) [3, 4], in the safety and security area e.g. for detection of explosives [5, 6] and in the field of first responders e.g. for detection of toxic industrial compounds (TIC) [7–9] are found in literature. First instruments were developed in early 1970, mainly for military applications [1]. With the end of the cold war the IMS technology became applicable also in other fields. Most of the instruments had one thing in common: a longitudinal drift region in the form of a tube. By applying a uniform potential gradient along the drift tube, a homogeneous electric field and by this means a constant accelerating force on charged particles is established. Under the influence of an ionization source target molecules are ionized and introduced into the drift tube. Most drift tubes consist of a reaction and a drift region separated by a shutter grid. The drift region is terminated by the detector. In the reaction region various charge transfer reactions take place forming so called product ions, which are injected into the drift region by a pulse like opening of the shutter grid. Under the influence of the electric field the product ions move towards the end of the drift region and are separated according to differences in ion mobility. The ions are recorded by a Faraday-cup detector. A major characteristic of this ion separation technique is the relatively weak strength of the applied electric field which is in the range from 150 to 300 V/cm. A details description can be found elsewhere[1, 2]. In 1991 the first papers appeared
132
describing an approach for separating ions according to their mobility differences in high and weak electric fields [10–12]. It was already known from Townsend and others that the dependence of the mobility coefficient from the electric field strength is more distinctive at higher electric field strength [1]. Based on this fact a new type of ion separation technology was introduced under the name, field asymmetric ion mobility spectrometry, (FAIMS) or, differential mobility spectrometry, (DMS). The basis is a separation in a high frequency asymmetric electric field. At frequencies up to 1 MHz the electric field strength can reach values of up to 30000 V/cm. Besides a cylindrical setup a planar setup is also known [13, 14]. The ions are transported by means of a carrier gas through a narrow gap bounded by two electrodes. The detector is located at the end of the channel. The asymmetric potential waveform is applied to the two electrodes placing the ions under the influence of a transversal electric field. The ions undergo a deflection during transition through the channel. Depending on the mobility properties, the deflection can lead to a recombination of the ions at the boundaries or, for some ions with specific mobility, passing through the channel and reach the detector. By applying an additional compensation voltage (CV) to the electrodes the deflection of the ions can be affected and an ion filter is realized, so that by scanning the compensation voltage a differential mobility spectrum can be obtained. The purpose of this work is to compare both methods of ion separation regarding detection capabilities, resolving power and sensitivity.
Instrument systems The measurements were conducted using two commercially available IMS instruments. For evaluation of the time of flight IMS a Gas Detector Array 2 (GDA2, Airsense Analytics, Germany) was used (http://www.sionex.com/products/index. htm). For the evaluation of the DMS a Sionex Value Added Component (SVAC, Sionex, Bedford, MA, USA) was used (http://www.airsense.com/english/index2.html). The GDA2 incorporates a classical IMS with a drift tube with a drift length of about 6.2 cm. The average electrical field strength in the reaction region is about 100 V/cm and in the drift region about 310 V/cm. The shutter grid is a conventional Nielson-Bradbury gate. A radioactive 100 MBq Ni63 foil acts as the ionization source and is located at the beginning of the drift tube. The drift gas runs in a closed loop with a built in molecular sieve filter. The vapor sample is introduced into the ionization region of the drift tube via a 10 µm polydimethylsiloxane (PDMS) membrane inlet. The membrane inlet is heated up to T=60 °C, the drift tube runs at a temperature T=37.5 °C±2.5 K. In order to measure positive and negative ions, the system reverses the polarity of
Int. J. Ion Mobil. Spec. (2009) 12:131–137
the applied electric potential every 1.5 s. In this manner both ion species can be detected quasi-simultaneously. The obtained spectra are recorded by software. An identical membrane inlet, which is used for the IMS in the GDA2, was adapted to the commercial DMS system, which usually comes with a direct inlet. The membrane inlet was maintained at a temperature of T=60 °C. The DMS cell temperature was controlled at T=80 °C. Pressurized synthetic air additionally purified with a combined molecular sieve / active charcoal filter is used as the carrier gas for the DMS. The carrier gas flow was adjusted to 300 ml/min with a mass flow controller. The sweep range for the compensation voltage was adjusted to match the needs of the actual measured compound to achieve the best resolution. The compensation voltage does not exceed −43 V and 15 V. For each analyzed vapor sample the amplitude of the asymmetric potential waveform (RF voltage) was stepwise increased from 500 V to 1500 V and at each step a compensation voltage sweep was conducted. This way a two dimensional dispersion plot could be obtained. Because the ions are carried by the gas flow through the separation region, the DMS is able to detect positive and negative ions simultaneously. The obtained dispersion plots are recorded by software.
Experimental The test vapor samples were prepared either from certified gas cylinders or from the evaporation of the liquid phase in sampling bags. The evaporation of liquid substances in an enclosure of known volume is a well known procedure. A PET-Polyester-foil bag with a volume of 3.8 l is filled with clean air. A few µl of the liquid (target solvent) are injected using a micro filter syringe. Applying physical law for ideal gases the amount of liquid Vinj to be injected can be calculated using the formula 1: Vinj ¼
ci Mi Vair ri 82:1 T
ð1Þ
Were ci is the required concentration of the vapor sample in ppmv, Mi is the molar mass in g/mol, Vair is the volume of the bag in l, ρi is the density in g/cm3 and T is the temperature in K. After a few minutes the solvent is normally evaporated; this can be easily controlled by visual inspection. From this sample, further bags with lower concentrations can be obtained by a dilution procedure. With a gas tight syringe a portion of the higher concentrated sample is transferred to further bags filled with clean air. This method for creating diluted gases is easy to apply but it has a large uncertainty in the concentration. Errors in concentration up to 40% have to be considered, depending
Int. J. Ion Mobil. Spec. (2009) 12:131–137
133
on the number of dilution steps. This method is still ideal for qualitative and comparative measurements. The instruments were challenged each time with the identical bags with the vapor samples. The measurement signals were recorded after achieving a steady state signal. Positive and negative spectra were saved using the IMS software. The DMS was operated in scanning mode in order to obtain dispersion plots. A single measurement with the DMS takes 100 s. Table 1 summarizes the measured compounds and the corresponding concentrations. From the drift time, the length of the drift tube and the applied electric field it is possible to calculate the mobility of each compound. Because of the well known pressure and the temperature dependence of the mobility of the ions in a drift-tube, a reduced mobility is introduced which makes the results between instruments more comparable. The calculation of the mobility with the applied pressure and temperature compensation leads to the reduced mobility as shown in equation 2: K0 ¼
l 273 K P td E T 1013hPa
ð2Þ
(Fig. 2) reactant ion peaks were evaluated for three different flow conditions: 200 ml/min, 300 ml/min and 400 ml/min. The increase of the reactant ion peak height in the spectrum indicates that more ions were collected at the detector with increased flow. With the increased flow rate the total throughput per time of ions is also increased which leads to the observed behavior in the positive as well as in the negative spectra. A shift in the peak position was not observed. For investigation of the pressure dependence the relative pressure in the DMS was increased in four stages 0 mbar, 100 mbar, 200 mbar and 510 mbar. Again the position and the height of the positive (Fig. 3) and negative (Fig. 4) reactant ion peaks were evaluated. A shift in the peak position could be observed. With increased pressure, smaller voltages are needed to compensate the ion deflection. This indicates that the differences in the weak and high field mobilities are less under high pressure conditions. The shift is more evident in the negative spectrum (0 mbar, CV: −14.68 V; 510 mbar, CV: −7.01 V) than in the positive (0 mbar, CV: −4.65 V; 510 mbar, CV: −2.29 V). A significant change in the peak height was not observed. A method for compensation of the pressure dependence is suggested in [15], but was not applied here.
Results and discussion Measurement results of selected TICs All the test gases were measured consecutively with the drift-tube IMS and the DMS. Because the influences of pressure and flow rate changes on the DMS are not clear, the dependencies were investigated and shown in the next chapter. Flow and pressure dependence of a DMS To investigate the influence of flow and pressure changes and in order to find optimal working parameters the DMS was operated under various flow and pressure conditions. The total flow through the DMS was modified using the mass flow controller. The position and the height of the positive (Fig. 1) and negative
Table 1 Measured compounds and corresponding concentrations 1 2 3 4 5 6 7 8
In the first part of the measurements, highly volatile solvents with small molecular masses forming mainly positive product ions were analyzed. Acetone at a concentration of 1ppmv shows a clear signal in the positive spectrum of the classical drift tube IMS (Fig. 5). Besides the reactant ion peak (K0 =2.10), the protonated acetone dimer peak (K0 =1.82) is evident. Even at lower concentrations of acetone a protonated monomer peak cannot be observed in the spectrum. This implies that the protonated acetone monomer has a comparable reduced mobility similar to the reactant ions and an overlapping of both species occurs in the spectrum. The resolving power of the used IMS does not allow a separation of these two
Compound
Measured concentration
Acetone Benzene Cyanogen chloride diisopropyl methylphosphonate (DIMP) dipropylen glycol monomethyl ether (DPM) Hydrogen cyanide Sulfur dioxide Benzene, 2,4-diisocyanato-1-methyl- (TDI)
100ppbv, 300ppbv, 1ppmv, 30ppmv 10ppmv 150ppbv, 1ppmv 100ppbv, 200ppbv 10ppbv, 100ppbv 1ppmv, 10ppmv, 100ppmv 100ppbv, 1ppmv 20ppbv, 50ppbv, 100ppbv, 300ppbv
134
Int. J. Ion Mobil. Spec. (2009) 12:131–137 rel. pressure : 510 mbar
flow: 300 ml/min flow: 200 ml/min
rel. pressure : 200 mbar
Intensity / V
Intensity / V
flow: 400 ml/min
rel. pressure : 100 mbar rel. pressure : 0 mbar
Compensation voltage / V
Compensation voltage / V
Fig. 1 Flow dependence of the RIP species in the DMS. Overlay of positive spectra (RF=800 V) at different flow rates (200 ml/min, 300 ml/min, 400 ml/min)
Fig. 3 Pressure dependence of the RIP species in the DMS. Overlay of positive spectra (RF=800 V) at different relative pressures (0 mbar, 100 mbar, 200 mbar, 510 mbar)
species. By evaluating the response from the DMS to 1ppmv acetone (Fig. 6) at RF voltages greater than 900 V it is interesting to see that the protonated acetone monomer becomes distinguishable from the reactant ions. At the corresponding field strengths the variations in the differential mobility becomes evident. At this concentration the protonated acetone dimer is also present. Comparing the results from the used IMS and the DMS it is clear that the DMS is able to resolve species with comparable weak field reduced mobility similar to the reduced mobility of the reactant ion. The measurement results using benzene are confirming this conclusion. Regardless of the concentration there is no specific signal in the drift tube IMS. In contrast, in the DMS a response to benzene could be observed. In the second part of the measurements, samples forming mainly negative product ions were analyzed. The test of 1ppmv cyanogen chloride shows a clear signal in the negative spectrum of the IMS (Fig. 7). The peak with the reduced mobility of K0 =−2.52 characterizes the cyanide product ion species which is clearly separated from the reactant ion species with a reduced mobility of K0 =−2.28. The response of the DMS to the same sample also shows a
clear response (see Fig. 8). At RF voltages greater than 1200 V only the cyanogens chloride product ion appears. The reason for the disappearance of the reactant ion is that at higher RF voltages the reactant ion starts to decompose. It is also interesting to see that the position of the hydrogen cyanide product ion in the drift tube IMS is the same as the product ion of cyanogen chloride. This means that with a classical drift tube IMS it is not possible to distinguish the two substances. In comparison, the DMS spectrum for hydrogen cyanide is different from the spectrum of cyanogen chloride. It can be concluded that product ions with reduced mobilities comparable to the reduced mobility of the reactant ion peak can be better resolved in the DMS than in the IMS. This applies for positive product ions as well as for negative product ions. In order to evaluate the response of the detectors to compounds with lower mobility (generally higher molecular mass) another set of measurements was also performed. Mainly phosphor organic compounds and simulants of chemical warfare agents were used for these tests. For comparison diisopropyl methylphosphonate (DIMP) and dipropylen glycol monomethyl ether (DPM) were analyzed (Fig. 9). Both substances show a clear signal in the IMS (Fig. 9a and d). At the applied
400 ml/min
rel. pressure : 510 mbar
flow:
rel. pressure : 200 mbar
300 ml/min flow: 200 ml/min
Compensation voltage / V
Fig. 2 Flow dependence of the RIN species in the DMS. Overlay of negative spectra (RF=800 V) at different flow rates (200 ml/min, 300 ml/min, 400 ml/min)
Intensity / V
Intensity / V
flow:
rel. pressure : 100 mbar rel. pressure : 0 mbar
Compensation voltage / V
Fig. 4 Pressure dependence of the RIN species in the DMS. Overlay of negative spectra (RF=800 V) at different relative pressures (0 mbar, 100 mbar, 200 mbar, 510 mbar)
Int. J. Ion Mobil. Spec. (2009) 12:131–137
135
K0=2.52
K0=2.28 RIN
Intensity / pA
Intensity / pA
K0=2.10 RIP-H2O
K0=1.82
Drift time / ms
Drift time / ms
Fig. 5 Positive spectrum of acetone (1 ppmv) taken with the time of flight IMS
Fig. 7 Negative spectrum of cyanogen chloride (1ppmv) taken with a time of flight IMS
concentration of 100ppbv the protonated monomers and dimers are evident for both compounds. The reduced mobilities of the monomers (DIMP: K0 =1.54, DPM: K0 = 1.67) are clearly distinguishable as well as the reduced mobilities of the dimers (DIMP: K0 =1.08, DPM: K0 = 1.21). In the DMS this group of substances is generally characterized by small compensation voltages around zero.
In the dispersion plots for each substance the protonated monomers and dimers can be clearly seen (Fig. 9b and e). The corresponding spectra at a fixed RF voltage of 1200 V (Fig. 9c and f) show the peaks of the protonated monomers and dimers. It can be seen from the spectra that at this RF voltage an overlap of the dimers is evident. Both appear at the same compensation voltage of 2.7 V. But the protonated
DMS Acetone 1ppm Positive Spectra
a
DMS ClCN 1ppm Negative Spectra 600
800
RF [V]
RF [V]
600
1000 1200
800 1000 1200
1400 -16
-14
-12
-10
-8
-6
-4
-2
0
2
1400 -40
4
compensation voltage [V]
-35
-30
-25
-20
-15
-10
-5
0
5
compensation voltage [V] acetone monomer
RIP 3
ClCN prod. Ion
Acetone 1ppm Positive Spectra at RF=1000V
b -0.15
CV=-8.6V
2
Intensity [V]
Intensity [V]
b
acetone dimer
CV=-1.4V
1 0
RIN
ClCN 1ppm Negative Spectra at RF=1000V CV=-20.74V
-0.14 -0.13 -0.12 -0.11
-1 -16
-14
-12
-10
-8
-6
-4
-2
0
2
4
compensation voltage [V]
Fig. 6 a Dispersion plot of the positive ions of acetone (1ppmv) taken with the DMS. b Positive spectrum of acetone (1ppmv) taken with the DMS at RF=1000 V
-40
-35
-30
-25
-20
-15
-10
-5
0
5
compensation voltage [V]
Fig. 8 a Dispersion plot of the negative ions of cyanogens chloride taken with the DMS. b Negative spectrum of cyanogens chloride (1ppmv) taken with the DMS at RF=1000 V
136
Int. J. Ion Mobil. Spec. (2009) 12:131–137 DIMP 100ppb Positive Spectra
b
a
600
RF [V]
K0=2.10 RIP
800 1000
1400 -40
-35
-30
-25
-20
-15
-10
-5
0
5
10
15
compensation voltage [V]
K0 Mon=1.54 K0 Dim=1.08
c 0.25 Intensity [V]
Intensity / pA
1200
DIMP 100ppb Positive Spectra at RF=1200V
0.2
CV=-0.23V CV=2.7V
0.15 0.1 0.05
-40
-35
-30
-25
-20
-15
-10
-5
0
5
10
15
5
10
15
compensation voltage [V]
Drift time / ms
d
DPM 100ppb Positive Spectra
e RF [V]
600
1200 1400 -40
-35
-30
-25
-20
-15
-10
-5
0
compensation voltage [V]
K0 Mon=1.67 K0 Dim=1.21
f Intensity [V]
Intensity / pA
K0=2.10 RIP
800 1000
DPM 100ppb Positive Spectra at RF=1200V
0.25 0.2
CV=-1.4V
CV=2.7V
0.15 0.1 0.05
-40
-35
-30
-25
-20
-15
-10
-5
0
5
10
15
compensation voltage [V]
Drift time / ms
Fig. 9 Measurement signals 100 ppbv DIMP and 100 ppbv DPM taken with the time of flight IMS in comparison with the DMS. a 100 ppbv DIMP positive spectrum taken with the time of flight IMS. b 100 ppbv DIMP dispersion plot of the positive ions taken with the DMS c: Positive spectrum of 100 ppbv DIMP taken with the DMS at
RF=1200 V d 100ppbv DPM positive spectrum taken with the time of flight IMS. e 100 ppbv DPM dispersion plot of the positive ions taken with the DMS. f Positive spectrum of 100 ppbv DPM taken with the DMS at RF=1200 V
monomers are still distinguishable (DIMP: CV=−0.23 V, DPM: CV=−1.4 V). Extreme elevated concentration can lead to a saturation of the reaction region and to a total consumption of the reactant ions and even to the elimination of the protonated monomer from the mobility spectrum [1]. At these high concentrations, where only the dimers are present a clear distinction such as in the classical drift tube IMS is not possible anymore. From the presented results and from the measurements of more than 20 other TICs (not presented here) it can be concluded that the mobility differences between high and low field conditions of substances with higher molecular masses are not as distinct as the mobility differences of substances with lower molecular masses.
From the measurements it can be stated that the limit of detections of the two instruments used are in the same range. For instance the estimated limit of detection for benzene, 2,4-diisocyanato-1-methyl- (TDI) was 20ppbv for both instruments and 100ppbv for sulfur dioxide also for both instruments.
Conclusions The performance, especially the selectivity and sensitivity, of a DMS was compared to a classical tube IMS. Before starting the comparison the operation parameters of the DMS were optimized. Especially the flow and
Int. J. Ion Mobil. Spec. (2009) 12:131–137
pressure dependence of the DMS was investigated. The measurements show that higher flow rates lead to higher ion concentrations at the detector and consequently lower limits of detection. The pressure dependence of the DMS is more evident for the negative ions and should be accounted. The performances of the IMS and the DMS were tested by applying different vapors of toxic industrial compounds at different concentrations to the instruments. Regarding resolution; the DMS performs better for ions with small molecular masses (high mobilities) e.g. solvents. Especially for compounds with a reduced mobility comparable to that of the reactant ion the DMS is often able to show some separation, which is not achievable with the used IMS. The IMS performs better for ions with higher molecular masses (low mobilities) e.g. phosphor organic compounds. For these types of compounds a clear distinction can only be achieved in the IMS. Regarding the limit of detection (LOD) both instruments show similar results. Depending on the compound the LOD is in the range of ppbv to low ppmv. Acknowledgments The authors would like to acknowledge SIONEX Corporation for supporting this work by providing the DMS. Furthermore, the valuable support and the helpful discussions with Dr. Erkinjon Nazarov are mentioned thankfully.
137
References 1. Eiceman GA, Karpas Z (2005) Ion mobility spectrometry, 2nd edn. Tayler & Francis/CRC Press, Boca Raton 2. Stach J, Baumbach JI (2002) IJIMS 5 1:1–21 3. Asbury GR, Wu C, Siems WF, Hill HH (2000) Anal. Chim. Acta 404:273–283 4. Krebs M.D., Zapata A.M., Nazarov E.G., Miller R.A., Costa I.S., Sonenshein A.L., Davis C.E.; IEEE Sensors Journal, Vol. 5, No. 4, August 2005 5. Fricano L, Goledzinowski M, Jackson R, Kuja F, May L, Nacson S (2001) IJIMS 4 2:22–26 6. Marr AJ, Groves DM (2003) Int. Soc. Ion Mobility Spectrom 6 (2):59–62 7. Leonhardt JW, Rohrbeck W, Bensch H (2000) IJIMS 3 1:43–49 8. Kanu AB, Hill HH, Gribb MM, Walters RN (2007) J. Environ. Monit. 9:51–60 9. Walte A., Muenchmeyer W., Ungethuem B.; International Symposium on Ion Mobility Spectrometry, Finland Mikkeli, 22.7-27.7.2007 10. Gorshkov M.P.; Inventor’s Certificate of USSR 966583 (1982) G01N27/62. 11. Buryakov I.A., Krylov E.V., Soldatov V.P.; Inventor’s Certificate of USSR 1485808(1989) G01N27/62 12. Buryakov IA, Krylov EV, Makas’ AL, Nazarov EG, Pervukhin VV, Rasulev U Kh (1991) Tech. Phys. Lett 17:60 13. Buryakov IA, Krylov EV, Nazarov EG, Rasulev U Kh (1993) Int. J. Mass Spectrom. Ion Processes 128:143–148 14. Shvartsburg AA (2009) Differential ion mobility spectrometry. Tayler & Francis/CRC Press, Boca Raton 15. Nazarov EG, Coy SL, Krylov EV, Miller RA, Eiceman GA (2006) Anal. Chem. 78:7697