Chemical ionization mass spectral analysis of pinacolyl alcohol and development of derivatization method using p-tolyl isocyanate

Abstract

A simple and routine chemical ionization method (CI) providing molecular weight information is developed for the unambiguous identification of 3,3-dimethyl-2-butanol known as pinacolyl alcohol (PA), a potential precursor for prohibited chemical weapons such as Soman, a nerve agent. This method meets the CI criteria used for the evaluation of proficiency tests conducted by OPCW avoiding use of reference chemical. In addition, a derivatization method using p-tolyl isocyanate for the analysis of PA using GC-MS has also been developed for the identification. This derivatization method facilitates the easy analysis of PA at low concentration present in environmental matrices. The efficiency of the derivatization is >99% with 10 fold increase in the sensitivity of the GC-MS signal. The electron ionization mass spectral fragmentation of the p-tolyl isocyanate (PTI) derivatives of three isomeric alcohols is also established. The positive CI of PTI derivatives of these alcohols resulted in molecular ion information and meets the OPCW criteria for CI method of analysis. Further, the retention indices for all the studied alcohols after derivatization help the unambiguous identification of these compounds present in complex environmental matrices.

---

Introduction

Pinacolyl alcohol (PA) is a highly volatile compound among the list of Schedule 2 chemicals covered under Chemical Weapons Convention (CWC).¹ PA is a commercially available secondary alcohol and is used as a precursor for the synthesis of highly toxic nerve agents, such as soman. As a part of off-site analysis for screening of CWC related compounds, analysis of PA gains importance because of its volatility, which makes it an early eluting chemical in gas chromatography (GC). The identification of PA present in various sample matrices is almost impossible when tedious sample preparation procedures are employed and may be lost even during concentration of the sample aliquots. GC-MS is the most suitable technique for the identification of this compound provided a proper solvent delay time is set and without solvent peak tailing.¹ GC and GC-EIMS analysis of PA and its retention indices are reported in the literature.^1–4 Houriet et al.⁵ studied the negative ion chemical ionization of alcohols including PA under ion cyclotron resonance mass spectrometry. Reagent gases of different nature have been used to study chemical ionization of aliphatic alcohols.^7,8 Gulacar et al.⁹ studied the ammonia CI mass spectrometry of various alcohols but the study was not emphasized towards branched secondary alcohols. GC-CIMS is used as a complementary technique to GC-EIMS for the unambiguous identification of any test chemical in the official proficiency tests conducted by the Organization for Prohibition of Chemical Weapons (OPCW), The Netherlands. The analyte should give a molecular or quasimolecular ion with at least 10% of the base peak intensity in the corresponding CI mass spectrum.^1,6 If this criteria does not meet then the data should be compared to that obtained with corresponding reference chemical under similar conditions. The majority of the participating laboratories in the proficiency tests, so far have used methane and iso-butane as reagent gases for the GC-CIMS analysis of PA, and used the reference standard for the unambiguous identification of PA because the method did not give the molecular ion information as per the OPCW criteria for CI method of analysis. One of the participating laboratories used ammonia as the reagent gas in the 19th OPCW official proficiency test, but the results were not known. To the best of our knowledge no reports are available on the chemical ionization studies of PA present in various sample matrices for routine and unambiguous identification, providing molecular or quasimolecular ion.

PA is an early eluting, highly volatile alcohol and directly amenable for GC and GC-MS analysis. But the analysis of PA is considered to be difficult particularly at low sample concentration and when sample preparation steps are involved. It is usually present in organic or aqueous environments. If PA is present in aqueous samples extraction with a suitable organic solvent and removal of the solvent to achieve good concentration are necessary. Removal of the volatile solvent is usually carried out by nitrogen purging. PA is mostly missed during these typical sample preparation steps and in particular at low analyte concentration.

Isocyanates are organic molecules characterized by an NCO moiety and are known to be highly reactive. The stabilization of the NCO group during workplace sampling is necessary for their subsequent laboratory analysis. These isocyanates are known to be highly reactive and selective towards water, alcohols, amines etc., and the reaction proceeds to completion within a short time even at low temperature. Urethane formation of alcohol with an isocyanate is an effective derivatization method for alcohols and has been used extensively for GC and HPLC analysis for the estimation of isocyanates present in the environment.¹⁰

R-NCO + R¹-OH → RNHCO-OR¹

Several isocyanates namely phenyl isocyanate, hexamethylene-1,6-diisocyanate (HDI), 4,4′-methylene bis (phenyl isocyanate) (MDI) etc., and their substitutes are estimated as their amine or alcohol derivatives.^11–17

The EI mass spectrum of trimethylsilyl (TMS) derivative of PA is available in the commercial mass spectral library,¹⁸ and derivatization of PA with BSTFA (N,O-bis (trimethylsilyl) trifluoro acetamide) is mentioned.¹⁹ But to the best of our knowledge, no experimental details are available in literature on the silylation of PA with BSTFA. To avoid the loss of PA during sample preparation and/or missing it during GC-MS analysis, a new derivatization method for PA is developed using p-tolyl isocyanate as the reagent. There are a number of mono or di-isocyanates, which can be used as reagents for this purpose.¹¹ Diisocyanates usually produce multiple products during reaction and typical reaction conditions such as reagent concentration and temperature etc. have to be maintained to achieve the required product selectively. Among the available mono isocyanates, p-tolyl isocyanate (PTI) was selected for the derivatization of PA, but in principle any isocyanate can be used for this purpose.

Experimental section

Pinacolyl alcohol (PA) was synthesized using a standard synthetic procedure²⁰ and the starting materials, acetaldehyde, and magnesium turnings were purchased from Aldrich, USA and S.D. Fine Chemicals, India, respectively. Dichloromethane, hexanol and 2-hexanol (HPLC grade) were purchased from E-Merck, India. The reagent gases, 10% ammonia balance helium (Spec Gases, India), deuterated ammonia (Stohler Isotope Chemicals, USA), methane (ECM Special Gases, UK) and iso-butane (Aldrich, USA) were commercially available. Pure helium was obtained from BOC India Ltd., India. A stock solution of 1000 ppm PA was prepared in dichloromethane and further diluted to 100 ppm with the same solvent. The chemical structure of the PA was characterized by mass spectral data.

The GC-MS analysis was carried out on an Agilent 6890N gas chromatograph (Agilent Technologies, USA) equipped with a model 5973i mass selective detector. A CP Sil 8 CB (Varian, The Netherlands) capillary column (30 m length, 0.25 mm i.d, and 0.25 μm film thickness) was used. For PA analysis, the oven was programmed from an initial temperature of 60 °C (2 min) to the final temperature of 120 °C at the rate of 10 °C min⁻¹, and the final temperature was held for 2 min. For all PTI derivatives, the oven was programmed from an initial temperature of 50 °C (2 min) to the final temperature of 280 °C at the rate of 10 °C min⁻¹, and the final temperature was held for 5 min. Helium at the rate of 1 mL min⁻¹ was used as the carrier gas under constant flow mode. The inlet and interface temperatures were kept at 150 °C and 200 °C, respectively. The ion source and quadrupole temperatures were kept at 230 °C and 150 °C, respectively for EI mode and 180 °C and 150 °C, respectively for CI mode. The MS was scanned from 29 to 600 Da for EI and 60 to 600 Da for CI mode. The MS was scanned from 70 to 600 Da when deuterated ammonia was used as the reagent gas. Electron energy 95.7 eV was used for positive chemical ionization (PCI) of various reagent gases. The reagent gas pressures were maintained at a regulator pressure of 200–250 kPa with a set flow of 10% of the total flow allowed in to the source for CI. 1 μl of 100 ppm pinacolyl alcohol solution was injected under split injection mode at a split ratio of 10:1.

Derivatization procedure

200 μl of 100 ppm solution of PA in dichloromethane was taken and 2 μl of p-tolyl isocyanate reagent was directly added and kept at 60 °C for 15 min. Cooled to room temperature and 2 ml of methanol was added to deactivate the excess reagent. 1 μl of this solution was directly injected in to GC-MS.

Calculation of retention indices

The retention indices were calculated using Kovat's formula²¹ with respect to standard n-alkane hydrocarbons (C₆–C₃₀). For this purpose, a solution containing 50 ppm each of hydrocarbons in hexane was injected into the GC-MS. The experiments were repeated three times to calculate the RIs. The calculated RIs for the alcohols after derivatization are summarized in Table 1.

Table 1 Partial EI-MS data and RIs for PTI derivatives of compounds 1, 2 and 3

PTI derivative of Compound No.	RI	Relative abundance (%)						Other ions, m/z
		m/z	m/z	m/z	m/z	m/z	m/z
		235	220	151	134	107	85
1	1789 ± 0.2	41	1	60	25	100	51	39, 41, 43, 45, 55, 57, 65, 69, 77, 79, 87, 91, 104, 106, 120, 132, 133, 148.
2	1954 ± 0.2	62	—	48	23	100	3	39, 41, 43, 55, 57, 65, 77, 79, 91, 104, 106, 120, 132, 133, 148, 162.
3	1858 ± 0.2	44	—	100	22	85	11	39, 41, 43, 45, 55, 57, 65, 77, 79, 91, 104, 106, 132, 133, 148, 162.

---

Results and discussion

The EI mass spectrum of the synthesized PA matched well with the reported spectrum.¹⁸ The spectrum contained weak molecular ion (M⁺˙) at m/z 102 (∼ 2%) and a more stable carbocation (CH₃)₃C⁺ at m/z 57 as the base peak, which resulted from the molecular ion (M⁺˙). The spectrum also included an abundant ion at m/z 45 (i.e., CH₃CHOH⁺) that is characteristic of secondary alcohols.² Other ions appeared at m/z 87 and m/z 69, resulting from the loss of ˙CH₃ from the molecular ion and subsequent loss of water from the ion at m/z 87.

CI spectra are known to depend on source temperature, reagent gas selected and its pressure and energy of the electron beam. To compare the role of the reagent gas in achieving molecular ion information all the chemical ionization experiments were conducted under similar optimized conditions (see experimental section). The sensitivity of PA was found to be very poor in negative ion mode compared to positive ion mode with the reagent gases used in the present study. Hence, only the positive chemical ionization of PA using methane, iso-butane and ammonia as reagent gases was studied.

Positive chemical ionization of PA

The positive chemical ionization (PCI) mass spectrum of PA using methane as reagent gas is shown in Fig. 1. The PCI (CH₄) showed an ion at m/z 85 as the base peak corresponding to [MH-H₂O]⁺ ion, and a low abundant ion at m/z 101 (∼4%) corresponding to [M − H]⁺ ion. PCI (i-C₄H₁₀) also produced similar ions and the abundance of the ion at m/z 101 was very low (∼2%). However, the PCI (NH₃) provided excellent results with an abundant ammonium adduct, [M + NH₄]⁺ at m/z 120. The spectrum showed the ion at m/z 120 as the base peak, in addition to other characteristic ions at m/z 69, 85 and 102 (Fig. 2a).

Fig. 1 PCI mass spectrum of PA using CH₄ as the reagent gas.

Fig. 2 PCI mass spectra of PA using (a) NH₃ (b) ND₃ as the reagent gases.

It is known that the key ions found in the PCI (NH₃) of saturated secondary alcohols were due to adduct formation, substitution reaction and their mono dehydrogenation product.²² The adduct ions were known to be favoured at low temperatures for smaller molecules while the substitution ions were favoured for larger molecules at higher temperatures. In the present case, the ion at m/z 120, [M + NH₄]⁺, was formed through adduct formation of PA with ammonium ion. The ion at m/z 102 was formed due to substitution reaction and such kind of substitution reactions were known to take place from [M + NH₄]⁺ with the loss of H₂O molecule.²² The ions at m/z 85 and 69 corresponded to [MH-H₂O]⁺ and [MH-H₂O–CH₄]⁺, respectively. The [M- CH₃]⁺ ion at m/z 87 was also observed that could be due to EI contribution. Formation of the above-mentioned ions in PCI (NH₃) was further confirmed by PCI (ND₃) data. The PCI (ND₃) spectrum was shown in Fig. 2b. The ion at m/z 102 observed in the PCI (NH₃) spectrum was shifted to m/z 105 confirming the formation of this ion by ND₃ substitution reaction. The [M + NH₄]⁺ ion at m/z 120, which expected to shift to m/z 124 with ND₃ reagent, rather appeared at m/z 125. This suggests that –OH proton of PA also exchanged with deuterium. The ion at m/z 145 appeared in the high mass region correspond to [M + N₂D₇]⁺ after the exchange of –OH proton with deuterium. The shift of other ions was not observed abundantly in Fig. 2b, due to the difference in the reagent gas pressures.

As per the OPCW test protocols, the CI method of analysis can only be accepted as a complementary technique if the relative abundance of the molecular/quasimolecular ion is above 10%. Hence, we attempted to study the effect of reagent gas flow on the relative abundance of the observed quasimolecular ion (m/z 120). It was found that the relative abundance of the ion at m/z 120 increased from 12% to 100% when the reagent gas (ammonia) flow was increased from 4% to 16% of the total flow. Therefore, the CI data obtained at the optimized experimental conditions (see experimental section) meets the OPCW criteria. Consequently, this method keeps off the use of reference compound for comparison.

Though the PCI (NH₃) enables easy confirmation of the molecular weight of PA as per the OPCW criteria, it is possible to miss this chemical during sample preparation at the stage of concentration step. To some extent, this can be avoided by derivatizing selectively with a suitable reagent. The derivatization step can also avoid the escape of PA in the solvent delay during GC-MS analysis, if the derivatives elute at a later time. Hence, we developed isocyanates as the selective derivatization reagent for alcohols. Though any isocyanate could be used for derivatization, here we present the data using p-tolyl isocyanate (PTI) as the reagent.

We optimized the volume of PTI reagent required, the reaction temperature and time for better yields with minimum interference from side products. The optimized conditions are given in the experimental section. We found the efficiency of the derivatization, under the used experimental conditions, is >99%. The total ion chromatogram (TIC) obtained for the reaction mixture after PTI derivatization of PA (Fig. 3) shows cleanliness of the reaction, where there is no interference of derivative with side products and the peak due to reagent (after quenching with methanol). Based on the signals obtained from GC-MS analysis of PA and its PTI derivative, we found 10 fold increase in the signal for PTI derivative (data not shown). The EI mass spectrum of the PTI derivative of PA (Fig. 3) obtained by GC-MS analysis showed an abundant molecular ion (M⁺˙ >40%) and the overall fragmentation was relatively more when compared to the underivatized PA. The cleavages of C–N and C–O bond were the major fragmentation routes from the molecular ion. The ion at m/z 151 can be attributed to the McLafferty type rearrangement that is usually observed in carbonyl group containing compounds (Scheme 1).

Fig. 3 TIC of PTI derivatized PA (top) and its EI mass spectrum (bottom).

Scheme 1 The proposed fragmentation pattern for PTI derivatives of, (1) PA, (2) hexanol and (3) 2-hexanol.

Differentiation of PTI derivatized isomeric alcohols under EI

In the proficiency tests conducted by the OPCW, PA is often spiked in the test sample matrices along with other C₆ alcohols to create ambiguity in the analyses. But with the present derivatization method the unambiguous identification of PA among other isomeric alcohols is possible. In the present study, pinacolyl alcohol (1) and two other isomeric alcohols namely, hexanol (2) and 2-hexanol (3) (normal and branched, respectively) were selected, derivatized and analyzed to differentiate them from the analyte of interest, PA.

The EI mass spectra of PTI derivatized hexanol (2) and 2-hexanol (3) are shown in Fig. 4. Among the three derivatized isomeric alcohols, hexanol resulted in an abundant molecular ion at m/z 235 (62%), indicating the stability of the ion. The other two alcohols are branched and resulted in less stable molecular ions (less than 45%). The loss of ˙CH₃ from the molecular ion is only observed in the case of PA, but not with the other two alcohols. The ion at m/z 85 representing ⁺C₆H₁₃ is more abundant in the case of PA. The other fragment ions observed in the three derivatized alcohols are similar, but show difference in their relative abundances. The structure of the ion at m/z 134 is confirmed as ⁺C₈H₈NO by its measured mass (134.0606) using GC-EI high-resolution mass spectrometry (GC-EI-HRMS). The general fragmentation pattern is shown in Scheme 1 and the partial EIMS data for all three compounds is presented in Table 1. Though the PTI derivatives of the studied alcohols showed clear-cut differences in their EI spectra, retention indices (RIs) of the respective compounds are required for unambiguous identification. The RIs were calculated for all the PTI derivatives with respect to the standard n-alkane hydrocarbons (C₆ to C₃₀) and values are tabulated in Table 1. It can be noted that the RI value for the derivatized PA is very much distinct compared with the derivatized isomeric alcohols.

Fig. 4 EI mass spectra of PTI derivatives of hexanol and 2-hexanol.

Positive chemical ionization of PTI derivatives

The PCI of the studied PTI derivatives of isomeric alcohols was carried out using iso-butane as the reagent gas under optimized experimental conditions (see experimental section). All the compounds show abundant [M + H]⁺ ion (60–100%) at m/z 236, which also meets the OPCW criteria on CI method of analysis. These criteria could not be met with underivatized compounds with methane and iso-butane as the reagent gases routinely used in proficiency tests. But the present derivatization method utilising the routine PCI technique (with reagent gas such as iso-butane) provided the molecular ion information of the studied C₆ alcohols under PCI (Fig. 5). The ion at m/z 152 was the base peak in the PCI spectrum of PTI derivative of PA that resulted by McLafferty rearrangement from the molecular ion. This rearrangement process is characteristic of higher alkyl esters.²³ The abundance of the ion at m/z 152 is ∼30% and ∼1% in PTI derivative of 2-hexanol and hexanol, respectively. Thus formation of the ion at m/z 152 is diagnostic of PA among the isomeric alcohols studied.

Fig. 5 PCI mass spectra of PTI derivatives of PA, hexanol and 2-hexanol.

Conclusions

PA is a commercially available secondary alcohol and is used as a precursor for the synthesis of highly toxic nerve agents. The analysis of PA gains importance because of its volatility, which makes it an early eluting chemical in gas chromatography. The identification of PA present in various sample matrices is almost impossible when tedious sample preparation procedures are employed and may be lost even during concentration of the sample solutions. In the present study, a CI method using ammonia as reagent gas has been developed for the confirmation of the molecular mass which meets the criteria proposed by the OPCW. Thus the present PCI (NH₃) method can be used as a complementary technique to EIMS for the unambiguous identification of PA present in complex environmental matrices. A derivatization method has been developed for the first time for the derivatization of PA using p-tolyl isocyanate. The derivatization reaction is clean with no interfering side products observed in the GC-MS analysis. The efficiency of the derivatization was found to be >99% with a 10 fold increase in the sensitivity. One can easily differentiate PA from other alcohols based on their GC retention times as well as general fragmentation pattern obtained by EIMS. The calculated RIs for the studied compounds show clear differences and the presence of PA can be confirmed unambiguously.

Acknowledgements

The authors thank the Director, IICT for the facilities and encouragement.

References

1. M. Mesilaakso, in Chemical weapons convention chemical analysis, ed. M. Mesilaakso, Wiley, Chichester, England, 2005.
2. R. A. Friedel, J. L. Shultz and A. G. Sharkey, Anal. Chem., 1956, 28, 926–934.
3. L. Tullberg, I.-B. Peetre and B. E. F. Smith, J. Chromatogr., A, 1976, 120, 103–113.
4. E. R. J. Wils, Fresenius J. Anal. Chem., 1990, 338, 22–27.
5. R. Houriet, D. Stahl and F. J. Winkler, Environ. Health Perspect., 1980, 36, 63–68.
6. M. R. V. S. Murty, S. Prabhakar, V. V. S. Lakshmi, U. V. R. V. Saradhi, T. Jagadeshwar Reddy and M. Vairamani, Anal. Chem., 2001, 77, 3406–3410.
7. M. Vairamani, U. A. Mirza and R. Srinivas, Mass Spectrom. Rev., 1990, 9, 235–258.
8. M. Vairamani, K. V. Siva Kumar and G. K. Viswanadha Rao, Ind. J. Chem., 1992, 31B, 9–14.
9. F. O. Gulacar, F. Mermoud, F. J. Winkler and A. Buchs, Helv. Chim. Acta, 1984, 67, 488–501.
10. G. Chavez, B. Bravo, N. Pina, M. Arias, E. Vivas, F. Ysambertt, N. Marquez and A. Caceres, Talanta, 2004, 64, 1323–1328.
11. P. Tremblay, J. Lesage, C. Ostiguy and H. V. Tra, Analyst, 2003, 128, 142–149.
12. M. A. Pudovik, N. E. Krepysheva, D. I. Kharitonov and A. N. Pudovik, Russ. J. Gen. Chem., 2001, 71, 325–326.
13. N. Ostah, G. Lawson, S. Zafar, G. Harrington and J. Hicks, Analyst, 2000, 125, 111–114.
14. D. Karlsson, J. Dahlin, A. Marand, G. Skarping and M. Calene, Anal. Chim. Acta, 2005, 534, 263–269.
15. B.-H. Kim, Y.-K. Kim and J. H. Ok, J. Chromatogr., B: Biomed. Sci. Appl., 2001, 752, 173–177.
16. E. Vangronsveld and F. Mandel, Rapid Commun. Mass Spectrom., 2003, 17, 1685–1690.
17. D. Karlsson, M. Dalene and G. Skarping, Analyst, 1998, 123, 1507–1512.
18. NIST/EPA/NIH Mass Spectral Library, Version ( 2002).
19. R. M. Black and B. Muir, J. Chromatogr., A, 2003, 1000, 253–281.
20. A. I. Vogel, Vogel's practical organic chemistry, Longman Group Ltd, London, England, 1956.
21. E. Kovats, Helv. Chim. Acta, 1958, 41, 1915–1932.
22. J. B. Westmore and M. M. Alauddin, Mass Spectrom. Rev., 1986, 5, 381–465.
23. D. G. I. Kingston, J. T. Bursey and M. M. Bursey, Chem. Rev., 1974, 74, 215–242.

---