- Email: [email protected]
Adjusting mobility scales of ion mobility spectrometers using 2,6-DtBP as a reference compound
Talanta 76 (2008) 1218–1223
Contents lists available at ScienceDirect
Talanta journal homepage: www.elsevier.com/locate/talanta
Adjusting mobility scales of ion mobility spectrometers using 2, 6-DtBP as a reference compound ¨ A.-K. Viitanen a,∗ , T. Mauriala b , T. Mattila c , A. Adamov b,d , C.S. Pedersen e , J.M. Makel a¨ a , a f a b,d ¨ M. Marjamaki , A. Sysoev , J. Keskinen , T. Kotiaho a
Aerosol Physics Laboratory, Department of Physics, Tampere University of Technology, P.O. Box 692, FI-33101 Tampere, Finland Laboratory of Analytical Chemistry, Department of Chemistry, University of Helsinki, P.O. Box 55, FI-00014 University of Helsinki, Finland Environics Oy, P.O. Box 1199, FI-70211 Kuopio, Finland d Division of Pharmaceutical Chemistry, Faculty of Pharmacy, University of Helsinki, P.O. Box 56, FI-00014 University of Helsinki, Finland e Department of Chemistry, University of Copenhagen, Universitetsparken 5, DK-2100 Copenhagen Ø, Denmark f Moscow Engineering Physics Institute (State University), Kashirskoe sh. 31, 115409 Moscow, Russia b c
a r t i c l e
i n f o
Article history: Received 8 January 2008 Received in revised form 16 May 2008 Accepted 21 May 2008 Available online 29 May 2008 Keywords: IMS Time-of-flight 2,6-DtBP Reference compound Instrument comparison
a b s t r a c t Performance of several time-of-flight (TOF) type ion mobility spectrometers (IMS) was compared in a joint measurement campaign and their mobility scales were adjusted based on the measurements. A standard reference compound 2,6-di-tert butylpyridine (2,6-DtBP) was used to create a single peak ion mobility distribution with a known mobility value. The effective length of the drift tube of each device, considered here as an instrument constant, was determined based on the measurements. Sequentially, two multi-peaked test compounds, DMMP and DIMP, were used to verify the performance of the adjustment procedure in a wider mobility scale. By determining the effective drift tube lengths using 2,6-DtBP, agreement between the devices was achieved. The determination of effective drift tube lengths according to standard reference compound was found to be a good method for instrument inter-comparison. The comparison procedure, its benefits and shortcomings as well as dependency on accuracy of literature value are discussed along with the results. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Ion mobility spectrometry (IMS) is a rapid analytical technique to detect and identify gas phase compounds [1]. IMS was developed in 1970s [2] and since then, the most widespread applications based on ion mobility have been the chemical warfare detectors and airport security [1,3]. A number of other applications have been published, e.g. from the field of atmospheric studies [4,5], environmental analysis [6] and health care [7]. The IMS-device construction can be divided into two major categories. Aspiration condensers [8–11] sample the ions and separate them by mobility in an electric field perpendicular to the gas flow, by practical configuration of electrometer strips or sampling slits on the side wall of the mobility analyzer. Usually a term DMA (differential mobility analyzer) is used in this context (e.g. [12–14]). However, most of the instruments rely on the time-of-flight (TOF) method [1–7,15,16].
∗ Corresponding author. Tel.: +358 3 31152133; fax: +358 3 31152600. E-mail address: anna-kaisa.viitanen@tut.fi (A.-K. Viitanen). 0039-9140/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.talanta.2008.05.030
In the TOF-method the drift time of an ion swarm under influence of weak electric field is measured and the mobility of an ion can be determined according to K=
l , Etd
(1)
where K is the electrical mobility of an ion, E is the electric field, l is the length of the drift region and td is the drift time of the ion. The mobility K is commonly normalized to standard conditions to facilitate comparison between different measurements according to K0 = K
T0 p , T p0
(2)
where T is the gas temperature and p is the pressure. Normalization of the mobility value removes the effect of pressure and reduces the effect of temperature on the ion mobility. The normalization procedure, however, does not remove the effect of cluster formation nor does it remove the effects of internal parameters such as changes in the drift tube length caused by heat expansion, or inhomogeneities of the electric field [15,16]. Conditions, for one, affect on the nature of ions and hence the
A.-K. Viitanen et al. / Talanta 76 (2008) 1218–1223
1219
Fig. 1. The measurement setup. Sample introduction is implemented using either diffusion method (a) or injection method (b).
ion mobility. With decreasing temperature larger clusters are being formed (e.g. [17,18]). The degree of hydration depends on temperature and prevailing moisture concentration (e.g. [17]). Furthermore there are other kinds of defects that arise from real effects in the non-ideal instruments, e.g. diffusion of the ions in the drift chamber and electronic response time of the detector, which usually tend to skew the instrument function and lower the instrument resolution. One fundamental issue is the determination of the zero point of the time-of-flight discussed by Spangler [19]. Instrument testing, comparison measurements, etc. require an efficient method to be utilized to obtain reliable results on the performance of the device. In IMS, either internal or external references can be used. In the case of an internal reference, a known gaseous compound, forming well-defined ions, is injected together with the sample ions. External reference, on the contrary, is measured independently. Sample introduction in IMS applications can be implemented by ionizing evaporated compounds or using electrospray. Electrospray technique is used, e.g. in the case of non-volatile compounds such as tetraheptyl ammonium bromide (THABr) [12] and other tetra-alkyl halides [12,20], PAMAM-dendrimers [21] or large protein molecules [22]. The technique utilizing non-volatile liquid compounds in an ESI-source is actually widely used for aspiration type of IMS-instruments, reaching down to mobility values such as 0.1 cm2 /Vs, or even 0.01 cm2 /Vs, but non-volatile compounds from electrospray may also be analyzed in the high mobility range of 1.0–2.5 cm2 /Vs, conventionally covered in gas phase IMS studies (e.g. [20]). All in all, a good reference compound, resulting in known ions either from an ionized gaseous compound or from an ESI-unit, should satisfy a variety of properties such as minor dependency on temperature and humidity, minor tendency to form clusters and strong ion signal close to the mobility values of the compounds of interest [15,23]. Compared to other IMS techniques, due to good resolution of TOF the spectrum formed by the reference peak can be verified with high accuracy. For commercial IMS-devices based on time-of-flight technique the resolving power is in the scale of 20–60 [24]. However, even higher values have been obtained with custom-made devices (e.g. [25]).
Some of the chemicals suggested as standards for positive ion mode ion mobility spectrometry were reviewed by Eiceman et al. [15], who concluded the 2,6-di-tert butylpyridine (2,6-DtBP) to be the best candidate, and, instead of using Eqs. (1) and (2) to determine the reduced mobility, they calculated the mobility values with the help of reference compound based only on time-of-flight of a compound under interest according to K0 (unknown) t (standard) K(unknown) = = d K(standard) K0 (standard) td (unknown)
(3)
As a parallel method for the use of Eq. (3) in instrument comparison, the reference compound can also be used to adjust the mobility scale of the test instrument by determining a crucial parameter for it, such as ‘effective drift tube length’ or some another related quantity. The aim of this study is to explore a suitable procedure to adjust the mobility scales and to perform inter-comparison between different IMS-devices. For this a standard reference compound 2,6-di-tert butylpyridine discussed earlier is used. This gaseous compound is assumed to produce well defined ions with known reduced mobility value of 1.42 cm2 /Vs, which is an estimation from the Fig. 4 in Eiceman et al. [15]. To further test the different procedures in inter-comparison of different devices, two different compounds, dimethylmethylphosphonate (DMMP) and di-iso-propyl methyl phosphonate (DIMP) are chosen as test compounds. The chosen compounds provide relatively wide coverage of mobilities. 2. Experimental 2.1. Chemicals and sample introduction Chemicals used in this study were 2,6-di-tert butylpyridine (Sigma–Aldrich, Steinheim, Germany), dimethylmethylphosphonate (Fluka, Switzerland) and di-iso-propyl methyl phosphonate (Alfa-Aesar Johnson Matthey Company, Germany). Pentane (Labscan, Ireland) was used for dilution. The diffusion tube method was used for 2,6-DtBP and the injection method for DMMP and DIMP. In the diffusion tube method a
1220
A.-K. Viitanen et al. / Talanta 76 (2008) 1218–1223
dimer peaks were possible to identify from the measurement spectra, and kept constant through out the measurements.
2.2. Ion mobility spectrometers
Fig. 2. Reduced mobility spectra of 2,6-DtBP for all three instruments in the comparison, derived by the nominal tube length (dashed line) and by the effective drift tube length (solid line). In the case of RAID 1 the effective drift tube length given by the manufacturer is used as a nominal drift tube length. The literature value for 2,6-DtBP (1.42 cm2 /Vs) is marked with gray vertical line. The y-axis is normalized to highest signal value.
small amount of liquid sample compound was placed into a container. This container was placed into a gas washing bottle. When the sample flow was guided through the gas washing bottle, a steady concentration was achieved. In the injection method, the glass syringe was filled with the liquid phase sample. The sample was then injected through a septum into the sample flow. In cases of DMMP and DIMP the dilution of 0.1% in pentane was used and the injection was performed using 100 l syringe (Hamilton Bonaduz AG, Switzerland) operated with syringe pump (Cole-Parmer Instrument Company, Vernon Hills, USA). The sample introduction methods are illustrated in Fig. 1. The concentration was adjusted so that both protonated monomer and proton bound
For the measurements, three different IMS-devices all based on time-of-flight method were used. One of the devices was a commer¨ cial ion mobility spectrometer (RAID 1, Bruker Daltonics GmbH) and two were custom-made IMS-devices, available at University of Helsinki, developed in cooperation with Moscow Engineering Physics Institute (State University). The first one of those (IMS1/MS), was mounted in front of a commercial API 300 triple quadrupole mass spectrometer (Sciex Applied Biosystems, Canada). Mass spectrometer was operated in selected ion monitoring (SIM) mode (m/z 192 for 2,6-DtBP, m/z 125 and 249 for DMMP, m/z 181 and 361 for DIMP). The voltages in IMS1/MS were supplied by self-made power source which was connected to the PC through a serial COM port. The IMS1/MS instrument has been described with details by Sysoev et al. [26]. The second device, an identical ion mobility spectrometer (IMS2) equipped with Faraday cup detector operated as a standalone unit. The resolving power of the RAID 1 was in the range of 20–30, while for the custom-made instruments (IMS1/MS and IMS2) it was in the range of 50–70 for all compounds measured in this study. ¨ For data acquisition of RAID 1 a Bruker IMS NT 3.0 software was used. A spectrum was obtained every 5 s after accumulation of four individual spectra (approximately 30 ms each). In data processing a median of six sequential spectra was formed to produce a final spectrum to be used in results. The IMS1/MS was operated in a dualgate mode [26]. Delay time was scanned in 50 s increments with the dwell time of 1 s. Maximum separation time was 50 ms. Gate opening times for both gates were 0.3 ms. The data was collected by Analyst 1.4.2 software (Sciex Applied Biosystems, Canada) and
Fig. 3. Mobility spectra of DMMP (a–b) and DIMP (c–d). Measurements with low vapor concentration above and with high vapor concentration below. Mobility distributions have been calculated using the nominal drift tube length (dashed lines) and the effective drift tube length corresponding to the reference value 1.42 cm2 /Vs for 2,6-DtBP (solid lines). In the case of RAID 1 the effective drift tube length given by the manufacturer is used as a nominal drift tube length. The y-axis is normalized to highest signal value.
A.-K. Viitanen et al. / Talanta 76 (2008) 1218–1223
1221
Table 1 Operation parameters of RAID 1, IMS1/MS and IMS2 during the measurements
Drift gas Gas flow Pressure Ionization method Drift voltage Drift tube length Electric field Pulse width a
RAID 1
IMS1/MS
IMS2
Air 0.4 l/min Ambient 63 Ni (15 mCi) 1465.4 V 5.76 cma 254 V/cm 0.3 ms
Air 1 l/min Ambient Corona 3508 V 13.3 cm 264 V/cm 0.3 ms
Air 1.7 l/min Ambient Corona 3517 V 13.85 cm 254 V/cm 0.3 ms
Effective drift tube length given by the manufacturer.
treated in MS Excel. The IMS1 was controlled by self-made software designed in LabVIEW (National Instruments, Austin, US). The IMS2 was operated in single gate mode [25]. The first gate opening time was 0.3 ms. The second ion gate was kept open at all times. Mobility distributions were recorded for 50 ms with repetition rate of 20 Hz. A 2000 transients were combined to obtain the analytical signal. With the IMS2 data collection and treatment was done using a selfmade LabVIEW program. The operation parameters of the devices are listed in the Table 1. 2.3. Measurement setup The measurement setup is presented in Fig. 1. In order to exclude all the external parameters possibly effecting on the measurements and the actual comparison of the devices, the measurements were performed in a joint campaign at a same laboratory. To synchronize the operating conditions, all the devices sampled from the same gas line simultaneously. The electric field strength of all three devices was adjusted to be close to each other. Ion pulse width of 0.3 ms was used in every device. The measurements were carried out in open cell conditions thus, the membrane normally used in the sample introduction system of RAID 1, was removed. In order to allow close electric fields in all three TOF-instruments, both IMS1/MS and IMS2 operated at reduced drift voltages (3508 V for IMS1/MS, and 3517 V for IMS2 instead normal operating drift voltage of 5040 V), though decrease of drift voltage resulted on significant decrease of resolving power reaching 100 at normal operation conditions for both instruments. Purified and dried (RH <1%) pressurized air was used as a carrier and drift gas. The carrier gas was divided into a sample and dilution flows. Gas flows were controlled with mass flow controller (Tele¨ a¨ dyne Hastings Instruments, Hampton, US) and rotameters (Kytol Instruments, Muurame, Finland). Teflon tubing was used to connect the parts. The instruments were not heated. The laboratory room was air conditioned, so that the temperature during measurements was constant ∼25 ◦ C. This temperature was used when calculating reduced mobilities for IMS1/MS and IMS2. Both of these devices are constructed so that the electronics will not heat up the drift tube substantially. Additionally, during the measurements, the temperature of in- and outgoing gas lines were measured and verified to be in line with the room temperature. For RAID 1, temperature was controlled using internal thermo sensor inside the drift tube wall. Pressure was monitored at the SMEAR III station of University of Helsinki, 51 m above sea level. The pressure varied between 1001 and 1012 mbar during the measurements. Pressure values used in calculations were obtained by determining the mean pressure value from the time period of the measurements. This value was used for calculating the reduced mobility of IMS1/MS and IMS2 measurements. For RAID 1 a pressure value measured by the device automatically was used.
Fig. 4. (a–c) Comparisons of measured and adjusted mobility values of 2,6-DtBP, DMMP and DIMP measured by each instrument pair (a) RAID 1-IMS1/MS; (b) RAID 1IMS2; (c) IMS1/MS-IMS2. White diamonds have been derived using nominal values for drift tube length and the black diamonds have been obtained using effective tube length that corresponds the 2,6-DtBP mobility 1.42 cm2 /Vs.
3. Results and discussion In principle, the reduced mobility values for this study could be calculated using three different procedures. In traditional method, a reduced mobility value for each compound is calculated from the original measured value using Eqs. (1) and (2) and physical parameters for each instrument, listed in Table 1. This procedure will give a straightforward result for the ion mobilities measured. In the second procedure, suggested by Eiceman et al. [15], Eq. (3) is used.
1222
A.-K. Viitanen et al. / Talanta 76 (2008) 1218–1223
There, the mobility value for a reference compound is used to obtain mobilites for the test compounds, and only the drift times of the reference and the test compounds in the same conditions are needed. This method gives straightforward values for reduced mobilities but does not provide any additional information about the actual instrument. The third method is to use also the known reduced mobility value for a reference compound, but tune the instrument parameters such as those presented in Table 1 according to that. In practice, this method assumes that some of the values in Table 1 may not actually be the real effective parameters of the device, but rather parameters based on the physical construction of the device. For instance, the electric field may not be fully homogeneous throughout the whole drift tube, and so the field strength derived from measured voltage and measured drift tube length, may not give a fully correct result. Therefore it is more practical to come up with an ‘effective’ value either for the field strength or the drift tube length. A feasible value for drift tube length is needed, e.g. in aerosol measurements, where the mobility distribution is derived from time-of-flight spectrum by help of instrument functions (e.g. [14]). Here, the effective drift tube length to be used as an instrument constant of each device is determined. First, the actual mobility value in measurement conditions is calculated and reduced to normal conditions according to Eqs. (1) and (2). This reduced mobility is then compared with the literature value thus, the effective drift tube length can be determined. Then, the effective drift tube length is used as a solid parameter to derive the mobilities for the test compound ions. From this analysis we obtain both an effective drift tube length for each instrument in the inter-comparison, plus the mobility values for the test compounds. This method is dependent only on the value used as a reduced mobility of the reference compound and therefore it should be known with great accuracy. Here, the reduced mobility of 2,6-DtBP (1.42 cm2 /Vs) given by Eiceman et al. [15] is used. The derived effective drift tube length values for RAID 1, IMS1/MS and IMS2 are 5.764, 12.900 and 13.621 cm, respectively. In the case of commercial RAID 1, though, one should recognize that the value for the drift tube length in Table 1 is the one given by the manufacturer. The actual physical drift tube length of RAID 1 was not available. In Fig. 2 in addition to the actual measured mobility distributions, also the adjusted distributions are shown as solid lines. Thus, Fig. 2 illustrates the shifting of the mobility distribution when the nominal drift tube length is replaced with effective drift tube length in case of reference compound 2,6-di-tert butylpyridine. For RAID 1 these two distributions are in practice the same, while the calibration had to be made based on the effective drift tube length given by the manufacturer instead of actual physical length of the drift tube. In the case of IMS1/MS and IMS2 the results of the calibration is more clearly seen. The further measurement spectra of the test compounds are shown in Fig. 3(a–d). From the measurement
results it can be seen, how adjustment procedure works in practice. The reduced mobility of the proton bound dimer peaks for both test compounds were in extremely good agreement between all the three devices after the adjustment. The peak formed by protonated monomers, however, showed some variation between the devices. The final comparisons of the reduced mobility values for the test compounds, obtained by each instrument using their effective drift tube lengths, are shown in Fig. 4(a–c). These figures illustrate that consistence between the instruments can be achieved when the adjustment procedure is used. When RAID 1 is compared to IMS1/MS and IMS2, indeed the strongest variation in mobility is seen for protonated monomers of the test compounds used here (DMMP and DIMP). But as we compare IMS1/MS and IMS2, variation in the protonated monomer values is no longer noticeable. Thus, by adjusting the length of the drift tubes consistency between IMS1/MS and IMS2 were achieved through out the measurement range. In Table 2 the reduced mobility values achieved after adjustment procedure are compared with both the reduced mobility values obtained using Eq. (3) and also with some literature values. When calculating the reduced mobility values using Eq. (3), we have used 2,6-DtBP as a standard mobility (1.42 cm2 /Vs) and corresponding drift times of each instrument as a standard drift times. Values obtained based on effective drift tube length correspond naturally to ones obtained using the method suggested by Eiceman et al. (Eq. (3)). Comparison of mobility values consist always some factors of uncertainties and hence literature values should be considered as indicative values. Here, the literature values seem to correspond to the measured values after adjustment within acceptable variation. Close examination of the spectral characteristics reveal some differences between the devices. For one, the reaction ion peak is slightly faster for RAID 1 than for IMS2, and the overall shape of the spectra produced by the test compounds DMMP and DIMP have minor differences. Also the reduced mobility values differ slightly between the devices, especially in the case of protonated monomers. The actual reduced mobility of the protonated monomer of DMMP is dependent on water vapor concentration [15,27]. Thus, the most probable cause for the different mobility values for the protonated monomers is different humidity inside the drift tubes. Other minor differences in the reduced mobilities and in the spectral characteristics are most probably caused by different construction of the devices, minor internal contaminations, and due to the fact that different ionization techniques [28,29] can give slightly different reduced mobility values. The most sensitive parameter in this calibration method is the value used for the reference compound, in this case for 2,6DtBP. At present, only one literature value by Eiceman et al. [15] is available. 2,6-DtBP is shown to have a minor temperature dependency in the scale of −10 to 250 ◦ C [15,18]. This is
Table 2 Reduced mobility values obtained in this study using effective drift tube length and Eiceman et al. [15] methods, and also corresponding literature values in cm2 /Vs 2,6-DtBP
DMMP1
DMMP2
1.42 1.42 1.42
1.74–1.80 1.76–1.80 1.80
1.38–1.39 1.39 1.39
1.82 1.63–1.81
1.42 1.45 1.42
DIMP1
DIMP2
Reference
1.48–1.53 1.49–1.53
1.08 1.08–1.09
1.54–1.55
1.09 1.13
This study, effective drift tube length methoda This study, Eiceman et al. [15] method (Eq. (3)) [15] ∼95 ◦ C (2,6-DtBP and DMMP2 ), ∼37 ◦ C (DMMP1 ), air, moisture 2.9 mg/m3 , values are estimations from graphical data [30] 100 ◦ C, air [31] Parameters not reported [27] Room temperature, air, water vapor concentration 6 ppm, 5 × 102 ppm and 2 × 102 ppm, values are estimations from graphical data
Subscripts 1 and 2 refer to protonated monomer and proton bound dimer, respectively. a The reduced mobility values have been obtained by determining the reduced mobility corresponding to maximum intensity value. When it was possible to detect the mobility peak from the spectra with low and high concentrations, the mobility value shown here is the average from these two.
A.-K. Viitanen et al. / Talanta 76 (2008) 1218–1223
in good agreement with experiments showing a high entropy change with hydration or dimerisation of 2,6-DtBP, due to the loss of rotational freedom of both the water molecule and the tert-butyl groups [32]. However, experimental ion mobility data on humidity behavior is not that well reported. Also measurement data with other carrier gases than air are lacking at present. Based on this study, 2,6-DtBP is well suited for reference compound in order to adjust the mobility scales. Before reaching the general approved status as an ion mobility reference compound, more measurement data and studies on 2,6-DtBP are needed in addition only to IMS methods based on time-of-flight technique. 4. Conclusion The mobility scales of three different IMS-devices, all based on time-of-flight technique, have been adjusted. The data required for the adjustment, was obtained in a joint measurement campaign. By equalizing the operational parameters, a comparable measurement setup was successfully created. The use of 2,6-DtBP as a standard compound enabled us to compare the devices by readjusting the mobility scale. 2,6-DtBP was well suited to a reference compound in our experience. By using DMMP and DIMP as a test compounds wider range in mobility scale was investigated, which confirmed the agreement between the instruments. The reduced mobility values also corresponded to the values available in the literature. Adjustment of the mobility scale is seen as an effective method when comparing different IMS-devices. Naturally, different conditions can still generate difference between the mobility distributions. Thus, the effect of conditions must be recognized. Based on this study, the use of 2,6-DtBP as a standard reference compound for IMS applications suggested by Eiceman et al. [15] is supported. For instrument comparison the determination of effective drift tube lengths using 2,6-DtBP as a reference compound is suggested. The obtained value can be then verified by measurements of other known compounds. Acknowledgements This study was supported by the Finnish Funding Agency for Technology and Innovation (Tekes), Environics Ltd., Orion Pharma, Paavo Ristola Consulting Engineers Ltd., and the National Bureau of Investigation (Finland) and the HERC project “Urban and rural air pollution—response of ecosystem and society (URPO)”. The authors would also like to acknowledge Finnish Defense Forces and
1223
Industrial Research Fund at Tampere University of Technology for ¨ was supported by the Academy of Finland, funding. Dr. Marjamaki project 109124. Division of Atmospheric Sciences of University of Helsinki is acknowledged for the pressure data. References [1] G.A. Eiceman, Z. Karpas, Ion Mobility Spectrometry, second ed., CRC Press, Boca Raton, 2005. [2] M.J. Cohen, F.W. Karasek, J. Chromatogr. Sci. 8 (1970) 330. [3] J.I. Baumbach, G.A. Eiceman, Appl. Spectrosc. 53 (1999) 338A. [4] K. Nagato, C.S. Kim, M. Adachi, K. Okuyama, J. Aerosol. Sci. 36 (2005) 1036. [5] K. Nagato, D.J. Tanner, H.R. Friedli, F.L. Eisele, J. Geophys. Res. 104 (1999) 3471. [6] L. Myles, T.P. Meyers, L. Robinson, Atmos. Environ. 40 (2006) 5745. [7] J.I. Baumbach, M. Westhoff, Spectrosc. Europe 18 (2006) 22. [8] J. Zeleny, Phys. Rev. 34 (1929) 310. [9] H.F. Tammet, The Aspiration Method for the Determination of Atmospheric-Ion Spectra, Translated from Russian, Published for the National Science Foundation, Washington, D.C. by the Israel Program for Scientific Translations, Jerusalem, 1970. [10] M. Zhang, A.S. Wexler, Int. J. Mass Spectrom. 258 (2006) 13. [11] S. Zimmermann, N. Abel, W. Baether, S. Barth, Sens. Actuators, B 125 (2007) 428. [12] S. Ude, J.F. de la Mora, J. Aerosol. Sci. 36 (2005) 1224. [13] T. Seto, K. Okuyama, L. de Juan, J.F. de la Mora, J. Chem. Phys. 107 (1997) 1576. ¨ ¨ V. Jokinen, T. Mattila, A. Ukkonen, J. Keskinen, J. Aerosol. Sci. 27 [14] J.M. Makel a, (1996) 175. [15] G.A. Eiceman, E.G. Nazarov, J.A. Stone, Anal. Chim. Acta 493 (2003) 185. [16] G.A. Eiceman, E.G. Nazarov, J.E. Rodriguez, J.A. Stone, Rev. Sci. Instrum. 72 (2001) 3610. [17] R.G. Ewing, G.A. Eiceman, J.A. Stone, Int. J. Mass Spectrom. 193 (1999) 57. ¨ ¨ M. Marjamaki, ¨ O. Anttalainen, J. Keskinen, [18] A.-K. Viitanen, T. Mattila, J.M. Makel a, Atmospheric Research (2008), doi:10.1016/j.atmos.res.2007.12.003. [19] G.E. Spangler, Anal. Chem. 65 (1993) 3010. [20] J. Viidanoja, A. Sysoev, A. Adamov, T. Kotiaho, Rapid Commun. Mass Spectrom. 19 (2005) 3051. [21] M. Imanaka, Y. Okada, K. Ehara, K. Takeuchi, J. Aerosol. Sci. 37 (2006) 1643. [22] S.L. Kaufman, J.W. Skogen, F.D. Dorman, F. Zarrin, Anal. Chem. 68 (1996) 1895. [23] Z. Karpas, Anal. Chem. 61 (1989) 684. [24] K. Cottingham, Anal. Chem. 75 (2003) 435A. [25] J. Laakia, T. Mauriala, A. Adamov, J. Viidanoja, V. Teplov, R.A. Ketola, R. Kostianen, A. Sysoev, T. Kotiaho, ISIMS2007, Mikkeli, Finland (2007). [26] A. Sysoev, A. Adamov, J. Viidanoja, R.A. Ketola, R. Kostiainen, T. Kotiaho, Rapid Commun. Mass Spectrom. 18 (2004) 3131. [27] R.G. Ewing, G.A. Eiceman, C.S. Harden, J.A. Stone, Int. J. Mass Spectrom. 255–256 (2006) 76. [28] H. Borsdorf, E.G. Nazarov, G.A. Eiceman, J. Am. Soc. Mass Spectrom. 13 (2002) 1078. [29] H. Borsdorf, E.G. Nazarov, G.A. Eiceman, Int. J. Mass Spectrom. 232 (2004) 117. [30] G.A. Eiceman, Y.-F. Wang, L. Garcia-Gonzalez, C.S. Harden, D.B. Shoff, Anal. Chim. Acta 306 (1995) 21. [31] P. Rearden, P.B. Harrington, Anal. Chim. Acta 545 (2005) 13. [32] M. Meot-Ner, L.W. Sieck, J. Am. Chem. Soc. 105 (1983) 2956.
Contents lists available at ScienceDirect
Talanta journal homepage: www.elsevier.com/locate/talanta
Adjusting mobility scales of ion mobility spectrometers using 2, 6-DtBP as a reference compound ¨ A.-K. Viitanen a,∗ , T. Mauriala b , T. Mattila c , A. Adamov b,d , C.S. Pedersen e , J.M. Makel a¨ a , a f a b,d ¨ M. Marjamaki , A. Sysoev , J. Keskinen , T. Kotiaho a
Aerosol Physics Laboratory, Department of Physics, Tampere University of Technology, P.O. Box 692, FI-33101 Tampere, Finland Laboratory of Analytical Chemistry, Department of Chemistry, University of Helsinki, P.O. Box 55, FI-00014 University of Helsinki, Finland Environics Oy, P.O. Box 1199, FI-70211 Kuopio, Finland d Division of Pharmaceutical Chemistry, Faculty of Pharmacy, University of Helsinki, P.O. Box 56, FI-00014 University of Helsinki, Finland e Department of Chemistry, University of Copenhagen, Universitetsparken 5, DK-2100 Copenhagen Ø, Denmark f Moscow Engineering Physics Institute (State University), Kashirskoe sh. 31, 115409 Moscow, Russia b c
a r t i c l e
i n f o
Article history: Received 8 January 2008 Received in revised form 16 May 2008 Accepted 21 May 2008 Available online 29 May 2008 Keywords: IMS Time-of-flight 2,6-DtBP Reference compound Instrument comparison
a b s t r a c t Performance of several time-of-flight (TOF) type ion mobility spectrometers (IMS) was compared in a joint measurement campaign and their mobility scales were adjusted based on the measurements. A standard reference compound 2,6-di-tert butylpyridine (2,6-DtBP) was used to create a single peak ion mobility distribution with a known mobility value. The effective length of the drift tube of each device, considered here as an instrument constant, was determined based on the measurements. Sequentially, two multi-peaked test compounds, DMMP and DIMP, were used to verify the performance of the adjustment procedure in a wider mobility scale. By determining the effective drift tube lengths using 2,6-DtBP, agreement between the devices was achieved. The determination of effective drift tube lengths according to standard reference compound was found to be a good method for instrument inter-comparison. The comparison procedure, its benefits and shortcomings as well as dependency on accuracy of literature value are discussed along with the results. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Ion mobility spectrometry (IMS) is a rapid analytical technique to detect and identify gas phase compounds [1]. IMS was developed in 1970s [2] and since then, the most widespread applications based on ion mobility have been the chemical warfare detectors and airport security [1,3]. A number of other applications have been published, e.g. from the field of atmospheric studies [4,5], environmental analysis [6] and health care [7]. The IMS-device construction can be divided into two major categories. Aspiration condensers [8–11] sample the ions and separate them by mobility in an electric field perpendicular to the gas flow, by practical configuration of electrometer strips or sampling slits on the side wall of the mobility analyzer. Usually a term DMA (differential mobility analyzer) is used in this context (e.g. [12–14]). However, most of the instruments rely on the time-of-flight (TOF) method [1–7,15,16].
∗ Corresponding author. Tel.: +358 3 31152133; fax: +358 3 31152600. E-mail address: anna-kaisa.viitanen@tut.fi (A.-K. Viitanen). 0039-9140/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.talanta.2008.05.030
In the TOF-method the drift time of an ion swarm under influence of weak electric field is measured and the mobility of an ion can be determined according to K=
l , Etd
(1)
where K is the electrical mobility of an ion, E is the electric field, l is the length of the drift region and td is the drift time of the ion. The mobility K is commonly normalized to standard conditions to facilitate comparison between different measurements according to K0 = K
T0 p , T p0
(2)
where T is the gas temperature and p is the pressure. Normalization of the mobility value removes the effect of pressure and reduces the effect of temperature on the ion mobility. The normalization procedure, however, does not remove the effect of cluster formation nor does it remove the effects of internal parameters such as changes in the drift tube length caused by heat expansion, or inhomogeneities of the electric field [15,16]. Conditions, for one, affect on the nature of ions and hence the
A.-K. Viitanen et al. / Talanta 76 (2008) 1218–1223
1219
Fig. 1. The measurement setup. Sample introduction is implemented using either diffusion method (a) or injection method (b).
ion mobility. With decreasing temperature larger clusters are being formed (e.g. [17,18]). The degree of hydration depends on temperature and prevailing moisture concentration (e.g. [17]). Furthermore there are other kinds of defects that arise from real effects in the non-ideal instruments, e.g. diffusion of the ions in the drift chamber and electronic response time of the detector, which usually tend to skew the instrument function and lower the instrument resolution. One fundamental issue is the determination of the zero point of the time-of-flight discussed by Spangler [19]. Instrument testing, comparison measurements, etc. require an efficient method to be utilized to obtain reliable results on the performance of the device. In IMS, either internal or external references can be used. In the case of an internal reference, a known gaseous compound, forming well-defined ions, is injected together with the sample ions. External reference, on the contrary, is measured independently. Sample introduction in IMS applications can be implemented by ionizing evaporated compounds or using electrospray. Electrospray technique is used, e.g. in the case of non-volatile compounds such as tetraheptyl ammonium bromide (THABr) [12] and other tetra-alkyl halides [12,20], PAMAM-dendrimers [21] or large protein molecules [22]. The technique utilizing non-volatile liquid compounds in an ESI-source is actually widely used for aspiration type of IMS-instruments, reaching down to mobility values such as 0.1 cm2 /Vs, or even 0.01 cm2 /Vs, but non-volatile compounds from electrospray may also be analyzed in the high mobility range of 1.0–2.5 cm2 /Vs, conventionally covered in gas phase IMS studies (e.g. [20]). All in all, a good reference compound, resulting in known ions either from an ionized gaseous compound or from an ESI-unit, should satisfy a variety of properties such as minor dependency on temperature and humidity, minor tendency to form clusters and strong ion signal close to the mobility values of the compounds of interest [15,23]. Compared to other IMS techniques, due to good resolution of TOF the spectrum formed by the reference peak can be verified with high accuracy. For commercial IMS-devices based on time-of-flight technique the resolving power is in the scale of 20–60 [24]. However, even higher values have been obtained with custom-made devices (e.g. [25]).
Some of the chemicals suggested as standards for positive ion mode ion mobility spectrometry were reviewed by Eiceman et al. [15], who concluded the 2,6-di-tert butylpyridine (2,6-DtBP) to be the best candidate, and, instead of using Eqs. (1) and (2) to determine the reduced mobility, they calculated the mobility values with the help of reference compound based only on time-of-flight of a compound under interest according to K0 (unknown) t (standard) K(unknown) = = d K(standard) K0 (standard) td (unknown)
(3)
As a parallel method for the use of Eq. (3) in instrument comparison, the reference compound can also be used to adjust the mobility scale of the test instrument by determining a crucial parameter for it, such as ‘effective drift tube length’ or some another related quantity. The aim of this study is to explore a suitable procedure to adjust the mobility scales and to perform inter-comparison between different IMS-devices. For this a standard reference compound 2,6-di-tert butylpyridine discussed earlier is used. This gaseous compound is assumed to produce well defined ions with known reduced mobility value of 1.42 cm2 /Vs, which is an estimation from the Fig. 4 in Eiceman et al. [15]. To further test the different procedures in inter-comparison of different devices, two different compounds, dimethylmethylphosphonate (DMMP) and di-iso-propyl methyl phosphonate (DIMP) are chosen as test compounds. The chosen compounds provide relatively wide coverage of mobilities. 2. Experimental 2.1. Chemicals and sample introduction Chemicals used in this study were 2,6-di-tert butylpyridine (Sigma–Aldrich, Steinheim, Germany), dimethylmethylphosphonate (Fluka, Switzerland) and di-iso-propyl methyl phosphonate (Alfa-Aesar Johnson Matthey Company, Germany). Pentane (Labscan, Ireland) was used for dilution. The diffusion tube method was used for 2,6-DtBP and the injection method for DMMP and DIMP. In the diffusion tube method a
1220
A.-K. Viitanen et al. / Talanta 76 (2008) 1218–1223
dimer peaks were possible to identify from the measurement spectra, and kept constant through out the measurements.
2.2. Ion mobility spectrometers
Fig. 2. Reduced mobility spectra of 2,6-DtBP for all three instruments in the comparison, derived by the nominal tube length (dashed line) and by the effective drift tube length (solid line). In the case of RAID 1 the effective drift tube length given by the manufacturer is used as a nominal drift tube length. The literature value for 2,6-DtBP (1.42 cm2 /Vs) is marked with gray vertical line. The y-axis is normalized to highest signal value.
small amount of liquid sample compound was placed into a container. This container was placed into a gas washing bottle. When the sample flow was guided through the gas washing bottle, a steady concentration was achieved. In the injection method, the glass syringe was filled with the liquid phase sample. The sample was then injected through a septum into the sample flow. In cases of DMMP and DIMP the dilution of 0.1% in pentane was used and the injection was performed using 100 l syringe (Hamilton Bonaduz AG, Switzerland) operated with syringe pump (Cole-Parmer Instrument Company, Vernon Hills, USA). The sample introduction methods are illustrated in Fig. 1. The concentration was adjusted so that both protonated monomer and proton bound
For the measurements, three different IMS-devices all based on time-of-flight method were used. One of the devices was a commer¨ cial ion mobility spectrometer (RAID 1, Bruker Daltonics GmbH) and two were custom-made IMS-devices, available at University of Helsinki, developed in cooperation with Moscow Engineering Physics Institute (State University). The first one of those (IMS1/MS), was mounted in front of a commercial API 300 triple quadrupole mass spectrometer (Sciex Applied Biosystems, Canada). Mass spectrometer was operated in selected ion monitoring (SIM) mode (m/z 192 for 2,6-DtBP, m/z 125 and 249 for DMMP, m/z 181 and 361 for DIMP). The voltages in IMS1/MS were supplied by self-made power source which was connected to the PC through a serial COM port. The IMS1/MS instrument has been described with details by Sysoev et al. [26]. The second device, an identical ion mobility spectrometer (IMS2) equipped with Faraday cup detector operated as a standalone unit. The resolving power of the RAID 1 was in the range of 20–30, while for the custom-made instruments (IMS1/MS and IMS2) it was in the range of 50–70 for all compounds measured in this study. ¨ For data acquisition of RAID 1 a Bruker IMS NT 3.0 software was used. A spectrum was obtained every 5 s after accumulation of four individual spectra (approximately 30 ms each). In data processing a median of six sequential spectra was formed to produce a final spectrum to be used in results. The IMS1/MS was operated in a dualgate mode [26]. Delay time was scanned in 50 s increments with the dwell time of 1 s. Maximum separation time was 50 ms. Gate opening times for both gates were 0.3 ms. The data was collected by Analyst 1.4.2 software (Sciex Applied Biosystems, Canada) and
Fig. 3. Mobility spectra of DMMP (a–b) and DIMP (c–d). Measurements with low vapor concentration above and with high vapor concentration below. Mobility distributions have been calculated using the nominal drift tube length (dashed lines) and the effective drift tube length corresponding to the reference value 1.42 cm2 /Vs for 2,6-DtBP (solid lines). In the case of RAID 1 the effective drift tube length given by the manufacturer is used as a nominal drift tube length. The y-axis is normalized to highest signal value.
A.-K. Viitanen et al. / Talanta 76 (2008) 1218–1223
1221
Table 1 Operation parameters of RAID 1, IMS1/MS and IMS2 during the measurements
Drift gas Gas flow Pressure Ionization method Drift voltage Drift tube length Electric field Pulse width a
RAID 1
IMS1/MS
IMS2
Air 0.4 l/min Ambient 63 Ni (15 mCi) 1465.4 V 5.76 cma 254 V/cm 0.3 ms
Air 1 l/min Ambient Corona 3508 V 13.3 cm 264 V/cm 0.3 ms
Air 1.7 l/min Ambient Corona 3517 V 13.85 cm 254 V/cm 0.3 ms
Effective drift tube length given by the manufacturer.
treated in MS Excel. The IMS1 was controlled by self-made software designed in LabVIEW (National Instruments, Austin, US). The IMS2 was operated in single gate mode [25]. The first gate opening time was 0.3 ms. The second ion gate was kept open at all times. Mobility distributions were recorded for 50 ms with repetition rate of 20 Hz. A 2000 transients were combined to obtain the analytical signal. With the IMS2 data collection and treatment was done using a selfmade LabVIEW program. The operation parameters of the devices are listed in the Table 1. 2.3. Measurement setup The measurement setup is presented in Fig. 1. In order to exclude all the external parameters possibly effecting on the measurements and the actual comparison of the devices, the measurements were performed in a joint campaign at a same laboratory. To synchronize the operating conditions, all the devices sampled from the same gas line simultaneously. The electric field strength of all three devices was adjusted to be close to each other. Ion pulse width of 0.3 ms was used in every device. The measurements were carried out in open cell conditions thus, the membrane normally used in the sample introduction system of RAID 1, was removed. In order to allow close electric fields in all three TOF-instruments, both IMS1/MS and IMS2 operated at reduced drift voltages (3508 V for IMS1/MS, and 3517 V for IMS2 instead normal operating drift voltage of 5040 V), though decrease of drift voltage resulted on significant decrease of resolving power reaching 100 at normal operation conditions for both instruments. Purified and dried (RH <1%) pressurized air was used as a carrier and drift gas. The carrier gas was divided into a sample and dilution flows. Gas flows were controlled with mass flow controller (Tele¨ a¨ dyne Hastings Instruments, Hampton, US) and rotameters (Kytol Instruments, Muurame, Finland). Teflon tubing was used to connect the parts. The instruments were not heated. The laboratory room was air conditioned, so that the temperature during measurements was constant ∼25 ◦ C. This temperature was used when calculating reduced mobilities for IMS1/MS and IMS2. Both of these devices are constructed so that the electronics will not heat up the drift tube substantially. Additionally, during the measurements, the temperature of in- and outgoing gas lines were measured and verified to be in line with the room temperature. For RAID 1, temperature was controlled using internal thermo sensor inside the drift tube wall. Pressure was monitored at the SMEAR III station of University of Helsinki, 51 m above sea level. The pressure varied between 1001 and 1012 mbar during the measurements. Pressure values used in calculations were obtained by determining the mean pressure value from the time period of the measurements. This value was used for calculating the reduced mobility of IMS1/MS and IMS2 measurements. For RAID 1 a pressure value measured by the device automatically was used.
Fig. 4. (a–c) Comparisons of measured and adjusted mobility values of 2,6-DtBP, DMMP and DIMP measured by each instrument pair (a) RAID 1-IMS1/MS; (b) RAID 1IMS2; (c) IMS1/MS-IMS2. White diamonds have been derived using nominal values for drift tube length and the black diamonds have been obtained using effective tube length that corresponds the 2,6-DtBP mobility 1.42 cm2 /Vs.
3. Results and discussion In principle, the reduced mobility values for this study could be calculated using three different procedures. In traditional method, a reduced mobility value for each compound is calculated from the original measured value using Eqs. (1) and (2) and physical parameters for each instrument, listed in Table 1. This procedure will give a straightforward result for the ion mobilities measured. In the second procedure, suggested by Eiceman et al. [15], Eq. (3) is used.
1222
A.-K. Viitanen et al. / Talanta 76 (2008) 1218–1223
There, the mobility value for a reference compound is used to obtain mobilites for the test compounds, and only the drift times of the reference and the test compounds in the same conditions are needed. This method gives straightforward values for reduced mobilities but does not provide any additional information about the actual instrument. The third method is to use also the known reduced mobility value for a reference compound, but tune the instrument parameters such as those presented in Table 1 according to that. In practice, this method assumes that some of the values in Table 1 may not actually be the real effective parameters of the device, but rather parameters based on the physical construction of the device. For instance, the electric field may not be fully homogeneous throughout the whole drift tube, and so the field strength derived from measured voltage and measured drift tube length, may not give a fully correct result. Therefore it is more practical to come up with an ‘effective’ value either for the field strength or the drift tube length. A feasible value for drift tube length is needed, e.g. in aerosol measurements, where the mobility distribution is derived from time-of-flight spectrum by help of instrument functions (e.g. [14]). Here, the effective drift tube length to be used as an instrument constant of each device is determined. First, the actual mobility value in measurement conditions is calculated and reduced to normal conditions according to Eqs. (1) and (2). This reduced mobility is then compared with the literature value thus, the effective drift tube length can be determined. Then, the effective drift tube length is used as a solid parameter to derive the mobilities for the test compound ions. From this analysis we obtain both an effective drift tube length for each instrument in the inter-comparison, plus the mobility values for the test compounds. This method is dependent only on the value used as a reduced mobility of the reference compound and therefore it should be known with great accuracy. Here, the reduced mobility of 2,6-DtBP (1.42 cm2 /Vs) given by Eiceman et al. [15] is used. The derived effective drift tube length values for RAID 1, IMS1/MS and IMS2 are 5.764, 12.900 and 13.621 cm, respectively. In the case of commercial RAID 1, though, one should recognize that the value for the drift tube length in Table 1 is the one given by the manufacturer. The actual physical drift tube length of RAID 1 was not available. In Fig. 2 in addition to the actual measured mobility distributions, also the adjusted distributions are shown as solid lines. Thus, Fig. 2 illustrates the shifting of the mobility distribution when the nominal drift tube length is replaced with effective drift tube length in case of reference compound 2,6-di-tert butylpyridine. For RAID 1 these two distributions are in practice the same, while the calibration had to be made based on the effective drift tube length given by the manufacturer instead of actual physical length of the drift tube. In the case of IMS1/MS and IMS2 the results of the calibration is more clearly seen. The further measurement spectra of the test compounds are shown in Fig. 3(a–d). From the measurement
results it can be seen, how adjustment procedure works in practice. The reduced mobility of the proton bound dimer peaks for both test compounds were in extremely good agreement between all the three devices after the adjustment. The peak formed by protonated monomers, however, showed some variation between the devices. The final comparisons of the reduced mobility values for the test compounds, obtained by each instrument using their effective drift tube lengths, are shown in Fig. 4(a–c). These figures illustrate that consistence between the instruments can be achieved when the adjustment procedure is used. When RAID 1 is compared to IMS1/MS and IMS2, indeed the strongest variation in mobility is seen for protonated monomers of the test compounds used here (DMMP and DIMP). But as we compare IMS1/MS and IMS2, variation in the protonated monomer values is no longer noticeable. Thus, by adjusting the length of the drift tubes consistency between IMS1/MS and IMS2 were achieved through out the measurement range. In Table 2 the reduced mobility values achieved after adjustment procedure are compared with both the reduced mobility values obtained using Eq. (3) and also with some literature values. When calculating the reduced mobility values using Eq. (3), we have used 2,6-DtBP as a standard mobility (1.42 cm2 /Vs) and corresponding drift times of each instrument as a standard drift times. Values obtained based on effective drift tube length correspond naturally to ones obtained using the method suggested by Eiceman et al. (Eq. (3)). Comparison of mobility values consist always some factors of uncertainties and hence literature values should be considered as indicative values. Here, the literature values seem to correspond to the measured values after adjustment within acceptable variation. Close examination of the spectral characteristics reveal some differences between the devices. For one, the reaction ion peak is slightly faster for RAID 1 than for IMS2, and the overall shape of the spectra produced by the test compounds DMMP and DIMP have minor differences. Also the reduced mobility values differ slightly between the devices, especially in the case of protonated monomers. The actual reduced mobility of the protonated monomer of DMMP is dependent on water vapor concentration [15,27]. Thus, the most probable cause for the different mobility values for the protonated monomers is different humidity inside the drift tubes. Other minor differences in the reduced mobilities and in the spectral characteristics are most probably caused by different construction of the devices, minor internal contaminations, and due to the fact that different ionization techniques [28,29] can give slightly different reduced mobility values. The most sensitive parameter in this calibration method is the value used for the reference compound, in this case for 2,6DtBP. At present, only one literature value by Eiceman et al. [15] is available. 2,6-DtBP is shown to have a minor temperature dependency in the scale of −10 to 250 ◦ C [15,18]. This is
Table 2 Reduced mobility values obtained in this study using effective drift tube length and Eiceman et al. [15] methods, and also corresponding literature values in cm2 /Vs 2,6-DtBP
DMMP1
DMMP2
1.42 1.42 1.42
1.74–1.80 1.76–1.80 1.80
1.38–1.39 1.39 1.39
1.82 1.63–1.81
1.42 1.45 1.42
DIMP1
DIMP2
Reference
1.48–1.53 1.49–1.53
1.08 1.08–1.09
1.54–1.55
1.09 1.13
This study, effective drift tube length methoda This study, Eiceman et al. [15] method (Eq. (3)) [15] ∼95 ◦ C (2,6-DtBP and DMMP2 ), ∼37 ◦ C (DMMP1 ), air, moisture 2.9 mg/m3 , values are estimations from graphical data [30] 100 ◦ C, air [31] Parameters not reported [27] Room temperature, air, water vapor concentration 6 ppm, 5 × 102 ppm and 2 × 102 ppm, values are estimations from graphical data
Subscripts 1 and 2 refer to protonated monomer and proton bound dimer, respectively. a The reduced mobility values have been obtained by determining the reduced mobility corresponding to maximum intensity value. When it was possible to detect the mobility peak from the spectra with low and high concentrations, the mobility value shown here is the average from these two.
A.-K. Viitanen et al. / Talanta 76 (2008) 1218–1223
in good agreement with experiments showing a high entropy change with hydration or dimerisation of 2,6-DtBP, due to the loss of rotational freedom of both the water molecule and the tert-butyl groups [32]. However, experimental ion mobility data on humidity behavior is not that well reported. Also measurement data with other carrier gases than air are lacking at present. Based on this study, 2,6-DtBP is well suited for reference compound in order to adjust the mobility scales. Before reaching the general approved status as an ion mobility reference compound, more measurement data and studies on 2,6-DtBP are needed in addition only to IMS methods based on time-of-flight technique. 4. Conclusion The mobility scales of three different IMS-devices, all based on time-of-flight technique, have been adjusted. The data required for the adjustment, was obtained in a joint measurement campaign. By equalizing the operational parameters, a comparable measurement setup was successfully created. The use of 2,6-DtBP as a standard compound enabled us to compare the devices by readjusting the mobility scale. 2,6-DtBP was well suited to a reference compound in our experience. By using DMMP and DIMP as a test compounds wider range in mobility scale was investigated, which confirmed the agreement between the instruments. The reduced mobility values also corresponded to the values available in the literature. Adjustment of the mobility scale is seen as an effective method when comparing different IMS-devices. Naturally, different conditions can still generate difference between the mobility distributions. Thus, the effect of conditions must be recognized. Based on this study, the use of 2,6-DtBP as a standard reference compound for IMS applications suggested by Eiceman et al. [15] is supported. For instrument comparison the determination of effective drift tube lengths using 2,6-DtBP as a reference compound is suggested. The obtained value can be then verified by measurements of other known compounds. Acknowledgements This study was supported by the Finnish Funding Agency for Technology and Innovation (Tekes), Environics Ltd., Orion Pharma, Paavo Ristola Consulting Engineers Ltd., and the National Bureau of Investigation (Finland) and the HERC project “Urban and rural air pollution—response of ecosystem and society (URPO)”. The authors would also like to acknowledge Finnish Defense Forces and
1223
Industrial Research Fund at Tampere University of Technology for ¨ was supported by the Academy of Finland, funding. Dr. Marjamaki project 109124. Division of Atmospheric Sciences of University of Helsinki is acknowledged for the pressure data. References [1] G.A. Eiceman, Z. Karpas, Ion Mobility Spectrometry, second ed., CRC Press, Boca Raton, 2005. [2] M.J. Cohen, F.W. Karasek, J. Chromatogr. Sci. 8 (1970) 330. [3] J.I. Baumbach, G.A. Eiceman, Appl. Spectrosc. 53 (1999) 338A. [4] K. Nagato, C.S. Kim, M. Adachi, K. Okuyama, J. Aerosol. Sci. 36 (2005) 1036. [5] K. Nagato, D.J. Tanner, H.R. Friedli, F.L. Eisele, J. Geophys. Res. 104 (1999) 3471. [6] L. Myles, T.P. Meyers, L. Robinson, Atmos. Environ. 40 (2006) 5745. [7] J.I. Baumbach, M. Westhoff, Spectrosc. Europe 18 (2006) 22. [8] J. Zeleny, Phys. Rev. 34 (1929) 310. [9] H.F. Tammet, The Aspiration Method for the Determination of Atmospheric-Ion Spectra, Translated from Russian, Published for the National Science Foundation, Washington, D.C. by the Israel Program for Scientific Translations, Jerusalem, 1970. [10] M. Zhang, A.S. Wexler, Int. J. Mass Spectrom. 258 (2006) 13. [11] S. Zimmermann, N. Abel, W. Baether, S. Barth, Sens. Actuators, B 125 (2007) 428. [12] S. Ude, J.F. de la Mora, J. Aerosol. Sci. 36 (2005) 1224. [13] T. Seto, K. Okuyama, L. de Juan, J.F. de la Mora, J. Chem. Phys. 107 (1997) 1576. ¨ ¨ V. Jokinen, T. Mattila, A. Ukkonen, J. Keskinen, J. Aerosol. Sci. 27 [14] J.M. Makel a, (1996) 175. [15] G.A. Eiceman, E.G. Nazarov, J.A. Stone, Anal. Chim. Acta 493 (2003) 185. [16] G.A. Eiceman, E.G. Nazarov, J.E. Rodriguez, J.A. Stone, Rev. Sci. Instrum. 72 (2001) 3610. [17] R.G. Ewing, G.A. Eiceman, J.A. Stone, Int. J. Mass Spectrom. 193 (1999) 57. ¨ ¨ M. Marjamaki, ¨ O. Anttalainen, J. Keskinen, [18] A.-K. Viitanen, T. Mattila, J.M. Makel a, Atmospheric Research (2008), doi:10.1016/j.atmos.res.2007.12.003. [19] G.E. Spangler, Anal. Chem. 65 (1993) 3010. [20] J. Viidanoja, A. Sysoev, A. Adamov, T. Kotiaho, Rapid Commun. Mass Spectrom. 19 (2005) 3051. [21] M. Imanaka, Y. Okada, K. Ehara, K. Takeuchi, J. Aerosol. Sci. 37 (2006) 1643. [22] S.L. Kaufman, J.W. Skogen, F.D. Dorman, F. Zarrin, Anal. Chem. 68 (1996) 1895. [23] Z. Karpas, Anal. Chem. 61 (1989) 684. [24] K. Cottingham, Anal. Chem. 75 (2003) 435A. [25] J. Laakia, T. Mauriala, A. Adamov, J. Viidanoja, V. Teplov, R.A. Ketola, R. Kostianen, A. Sysoev, T. Kotiaho, ISIMS2007, Mikkeli, Finland (2007). [26] A. Sysoev, A. Adamov, J. Viidanoja, R.A. Ketola, R. Kostiainen, T. Kotiaho, Rapid Commun. Mass Spectrom. 18 (2004) 3131. [27] R.G. Ewing, G.A. Eiceman, C.S. Harden, J.A. Stone, Int. J. Mass Spectrom. 255–256 (2006) 76. [28] H. Borsdorf, E.G. Nazarov, G.A. Eiceman, J. Am. Soc. Mass Spectrom. 13 (2002) 1078. [29] H. Borsdorf, E.G. Nazarov, G.A. Eiceman, Int. J. Mass Spectrom. 232 (2004) 117. [30] G.A. Eiceman, Y.-F. Wang, L. Garcia-Gonzalez, C.S. Harden, D.B. Shoff, Anal. Chim. Acta 306 (1995) 21. [31] P. Rearden, P.B. Harrington, Anal. Chim. Acta 545 (2005) 13. [32] M. Meot-Ner, L.W. Sieck, J. Am. Chem. Soc. 105 (1983) 2956.