macquarie university researchonline final version.pdf · 54 extraction for analysis by gas...
TRANSCRIPT
Macquarie University ResearchOnline
This is the author version of an article published as: Volk, Herbert; Fuentes, David; Fuerbach, Alexander; Miese, Christopher; Koehler, Wolfgang; Bärsch, Niko and Barcikowski, Stephan 2010 First on-line analysis of petroleum from single fluid inclusions using ultrafast laser ablation, Organic Geochemistry, Vol. 41, Issue 2, p.74-77
Access to the published version: http://dx.doi.org/10.1016/j.orggeochem.2009.05.006 Copyright: 2009 Elesevier
Volk et al single inclusion analysis_revised_Not Tracked.doc 20/09/2010 page 1 of 16
First on-line analysis of petroleum from single inclusion 1 using ultrafast laser ablation 2 3
Herbert Volk1*, David Fuentes1, Alexander Fuerbach2, Christopher Miese2, Wolfgang 4 Koehler3, Niko Bärsch4 and Stephan Barcikowski4 5 6 1) CSIRO Petroleum, North Ryde, Sydney, NSW 2113, Australia. 7 2) MQPhotonics Research Centre, Macquarie University, Sydney, NSW 2119, 8
Australia. 9 3) Femtolasers Produktions GmbH, Fernkornagasse 10, 1100 Wien, Austria. 10 4) Laser Zentrum Hannover e.V., Hollerithallee 8, 30419 Hannover, Germany. 11 12 Submitted to Organic Geochemistry, Special Issue: AOGC2008 13 14 *Corresponding Author, E-mail: [email protected], Telephone +61 2 9490 8954, 15 Facsimile +61 2 9490 8197 16 17
Abstract 18
For many years, geochemical analysis of petroleum from single fluid 19
inclusions has been a challenging objective for fluid inclusion studies. In 20
this study, individual petroleum inclusions have been selectively opened and 21
analysed, for the first time, by coupling an on-line femtosecond laser with a 22
gas chromatograph – mass spectrometer (GC-MS). 23
GC-MS chromatograms show straight-chained, branched and cyclic 24
alkanes and aromatic hydrocarbons with carbon numbers ranging from 4 25
(iso-butane) to 19 (pristane). The distribution of these compounds is similar 26
to that observed by on-line bulk crushing, and pyrolysis artefacts such as 27
alkenes and ketones were not detected. Hydrocarbons with higher carbon 28
numbers appear to have remained in the extraction chamber, a limitation 29
that may be overcome by improvements to the inlet system. 30
Volk et al single inclusion analysis_revised_Not Tracked.doc 20/09/2010 page 2 of 16
This pilot study proves that ultrashort laser pulses can be used to liberate 31
unaltered oil from individual inclusions, which opens exciting research 32
opportunities to deconvolute information on the nature of different 33
hydrocarbon palaeo-fluids from single minerals. 34
Keywords 35
Fluid inclusion; petroleum inclusion; femtosecond laser; laser ablation 36
37
Introduction 38
Petroleum fluid inclusions are tiny, usually <10 µm-sized vacuoles that are 39
filled with petroleum and their occurrence in rocks and minerals is well 40
documented (Burruss, 1981). They can often be found in quartz crystals and 41
sand grains of petroliferous basins, where they occur in healed fractures and 42
along crystal growth zones. The petroleum in these perfectly sealed vessels 43
provides a snap-shot of the fluid composition in the geological past, as it is 44
shielded from contamination and alteration that usually affects 45
hydrocarbons over time. Thus, unique information on the fluid evolution can 46
be obtained by analysing the chemical composition of the trapped 47
hydrocarbons. To date, the chemical composition of petroleum trapped in 48
fluid inclusions can be determined by inference from spectroscopic (e.g. 49
measurement of epifluorescence when excited with ultraviolet light, Raman 50
spectroscopy, Fourier-transform infrared spectroscopy) and 51
microthermometric methods (Munz, 2001) and by mechanical crushing or 52
thermal decrepitation of bulk samples followed by solvent or thermal 53
Volk et al single inclusion analysis_revised_Not Tracked.doc 20/09/2010 page 3 of 16
extraction for analysis by gas chromatography – mass spectrometry (GC-54
MS) (George et al., 2007). However, fluorescence microscopy and heating – 55
freezing experiments commonly indicate that these crystals contain several 56
fluid inclusion assemblages that hold information from different geological 57
events (e.g. Schubert et al., 2007). While spectroscopic methods, sometimes 58
combined with volumetric measurements and modelling (e.g. Thiéry et al., 59
2000), can constrain the composition of hydrocarbons in fluid inclusions 60
without loosing the spatial resolution provided by a microscope, detailed 61
molecular and isotopic data that hold the key to source and maturity of the 62
oil can so far only be obtained by using the crushing / GC-MS analysis 63
approach. However, as the samples are extracted in bulk, distinguishing the 64
geochemical signature of different fluid inclusion assemblages in the same 65
mineral is not possible. 66
The extraction of fluid inclusions using laser systems can potentially 67
bridge this gap. In some previous studies lasers have been used to extract oil 68
from fluid inclusions, yet these approaches were not selective enough to 69
differentiate individual inclusions. Greenwood et al. (1998) analysed 10-100 70
random inclusions using thermal decrepitation induced by an infrared 71
Nd:YAG laser beam with a wavelength of 1064 nm coupled on-line to a GC-72
MS system. Hode et al. (2006) extracted fluid inclusion oil in a vacuum 73
chamber using an infrared Er:YAG laser with a wavelength of 2940 nm and 74
condensation of this oil onto a cold finger. The condensed oil was 75
subsequently dissolved in hexane and analysed off-line by GC-MS. While 76
Volk et al single inclusion analysis_revised_Not Tracked.doc 20/09/2010 page 4 of 16
their approach could extract oil from a group of inclusions and individual 77
inclusions could be opened with this laser, it is not possible to analyse 78
volatile or gaseous hydrocarbons in this manner. Moreover, the infrared 79
wavelengths of both the Nd:YAG and Er:YAG lasers couple with the 80
petroleum trapped in the fluid inclusion rather than with the transparent 81
host mineral, such that the inclusions are explosively decrepitated. We 82
found decrepitation of selected inclusions using a Nd:YAG laser hard to 83
control and suspect that the decrepitation mechanism is likely to cause 84
other inclusions to open, as probably occurs in healed, yet still structurally 85
defective zones of the mineral. 86
In the present work, individual petroleum inclusions have been selectively 87
opened and analysed for the first time, by coupling a femtosecond laser on-88
line to a GC-MS instrument. 89
90
Experimental setup 91
The laser extraction GC-MS setup was a modified version of that used by 92
Greenwood et al. (1996). It comprises a Hewlett Packard 6890 gas 93
chromatograph (GC) interfaced to a Hewlett Packard 5973 mass selective 94
detector (electron energy 70 eV, electron multiplier 1200 V, source 95
temperature 250 ºC, 0.1 amu resolution) and an Olympus BX60M system 96
microscope with a custom-built laser chamber and inlet system. The 97
microscope was used with a long-working-distance objective (20x with 98
numerical aperture of 0.4). The Nd:YAG laser of this existing laser 99
Volk et al single inclusion analysis_revised_Not Tracked.doc 20/09/2010 page 5 of 16
micropyrolysis (LaPy) – GC-MS system was replaced by a high-energy 100
femtosecond oscillator system (Femtosource scientific XL 200, Femtolasers 101
GmbH). The pulses emitted by this laser have a spectral width of about 50 102
nm centred in the near-infrared around 800 nm. In contrast to continuous-103
wave radiation or q-switched (nanosecond) laser pulses, the energy in this 104
case is concentrated into an extremely short, i.e. only about 50 femtoseconds 105
long time interval. This results in an ultrahigh peak intensity at the laser 106
focus (>1014 W/cm2) that induces nonlinear (multiphoton) absorption of the 107
radiation, a process that is virtually independent on the actually centre 108
wavelength of the laser (Lenzner 1999, Martin et al., 2003). 109
The minimum energy that is required to exceed the ablation threshold and 110
to drill micro holes into quartz can be reduced if ultrashort laser pulses are 111
used. For pulses with a duration of 50 fs, the threshold fluence is only about 112
3 J/cm2 (Lenzner 1999). Combined with the fact that with the long-working-113
distance objective used in our setup the laser can be focussed to a minimum 114
spot diameter of about 2 µm, a pulse energy in the order of 100 nJ is 115
required. However, standard femtosecond oscillators (laser systems that are 116
relatively inexpensive, compact and user friendly) only offer pulse energies 117
of up to about 10 nJ. To our knowledge only femtosecond amplifier systems 118
have been employed for the ablation of transparent dielectric media so far. 119
Such laser systems, however, are expensive, complex and require a well 120
controlled laboratory environment, with minimal variations in temperature 121
and humidity, and have to be set-up on vibration-free optical tables. 122
Volk et al single inclusion analysis_revised_Not Tracked.doc 20/09/2010 page 6 of 16
Moreover, it has been shown that the required spatial resolution in the few-123
micrometer range can only be achieved in the ablation process if the pulse 124
energy is adjusted just above the actual ablation threshold of the material to 125
be processed (Gattass and Mazur, 2008). This implies that the output of 126
femtosecond amplifiers, typically in the mJ range, has to be strongly 127
attenuated for high precision micromachining, which greatly reduces the 128
penetration rate. 129
In this work we have chosen to use a novel and unique laser system that 130
avoids these drawbacks (Fig. 1). It can deliver pulse energies in the order of 131
hundreds of nanojoules while offering the same excellent performance in 132
terms of compactness and user friendliness as standard oscillators. By 133
introducing a Herriot-type multipass cell into the resonator of a standard 134
femtosecond oscillator, its repetition rate can be reduced by more than an 135
order of magnitude. At constant average output power, this results in a 136
substantially increased energy per laser pulse that can be sustained inside 137
the cavity by utilising an innovative dispersion management scheme 138
(Fuerbach et al., 2005). 139
The GC-MS was operated in full-scan mode (50-550 amu) and 140
chromatography was carried out using a DB-5 column (J&W, 25 m , 0.32 141
mm I.D, 0.52 �m film thickness) with helium as a carrier gas (1.5 mL/min). 142
The sample chamber was held at ca. 100°C, and helium at 100 mL/min 143
flushed petroleum liberated by the laser ablation process into a coiled nickel 144
loop immersed in liquid nitrogen. After 2 minutes of trapping and 145
Volk et al single inclusion analysis_revised_Not Tracked.doc 20/09/2010 page 7 of 16
cryofocussing, the helium flow through the nickel trap was reduced to 1 146
mL/min by switching a 6-port Valco valve, the cryotrap was removed and 147
the nickel trap heated to 320 ºC. Products were then re-focussed in a loop of 148
GC column immersed in liquid nitrogen. Once this trap was removed, the 149
GC oven was programmed from an initial temperature of 40 ºC (2 min. 150
hold), followed by heating at 4 ºC/min to 310 ºC (30 min. hold). 151
152
Sample 153
Our new analytical approach was trialled on a well-characterised 154
idiomorphic clear bipyramidal quartz crystal of ca. 3 mm size that was 155
isolated from a volcanic dyke cross-cutting organic-rich Silurian graptolite 156
shales of the Kopanina Formation at the quarry Kosov in the Barrandian 157
Basin (Czech Republic). Petroleum inclusions in this crystal are abundant 158
and can reach a large size of up to 100 µm in diameter. Inclusions are 159
dominated by abundant blue-fluorescing inclusions with some larger and 160
irregular-shaped inclusions showing a yellow fluorescence colour. More 161
detailed descriptions of the fluid inclusion assemblages in the test sample 162
(KQ7-Qz) and geochemical results from on-line bulk crushing are provided 163
in Volk (2000) and Volk et al. (2002). 164
165
Volk et al single inclusion analysis_revised_Not Tracked.doc 20/09/2010 page 8 of 16
Results 166
In our first attempt to selectively open a single inclusion we have 167
specifically targeted a small blue fluorescing (<10 µm sized) inclusion. We 168
focused laser pulses with an energy of 160 nJ and a pulse duration of about 169
50 fs at a repetition rate of 11 MHz onto the crystal, thereby ablating the 170
quartz and successfully opening the inclusion. This did not appear to affect 171
nearby inclusions. However, the amount of petroleum that was liberated 172
was too small for confident detection on our GC-MS system. 173
We then focused on a different and larger (> 50 µm) blue fluorescing 174
inclusion that is shown in Fig. 2a. Again, we extracted the petroleum from 175
this single inclusion without affecting adjacent inclusions, including one 176
with a more yellowish fluorescence colour deeper in the crystal (see Fig. 2b). 177
The extraction hole ablated into the quartz crystal had a diameter of about 2 178
µm (Fig. 2c). For this larger inclusion, the amount of extracted hydrocarbons 179
was sufficient to produce a clear signal on the GC-MS. The resulting 180
chromatograms show straight-chained, branched and cyclic alkanes and 181
aromatic hydrocarbons with carbon numbers ranging from 4 (iso-butane) to 182
19 (pristane) (Fig. 2d), while gases were not monitored with the chosen GC-183
MS scan rate. The distribution of low- to mid-weight molecular compounds 184
is similar to that observed by on-line bulk crushing (Volk et al., 2002), 185
however, C19+ compounds that were detected up to n-C31 in the on-line 186
crushing approach could not be detected using our laser approach. Similar 187
to previously reported laser extraction methods (e.g. Greenwood et al., 1998, 188
Volk et al single inclusion analysis_revised_Not Tracked.doc 20/09/2010 page 9 of 16
Hode et al., 2006) pyrolysis artefacts such as alkenes and ketones were also 189
not detected when using the pulsed laser system. This proves that the laser 190
pulses themselves do not alter the oil composition and do not introduce any 191
thermal artefacts. This can be explained by the fact that due to their 192
extremely high peak intensities, femtosecond laser pulses transfer their 193
energy to the sample via nonlinear optical processes rather then via thermal 194
absorption and thus alter the material by ‘cold ablation’ (Keller, 2003). More 195
detail on the distribution of low-molecular-weight hydrocarbons in the C5-196
C9 range is provided in Fig. 2e. 197
198
Outlook 199
It appears as if hydrocarbons with carbon numbers > 20 have remained in 200
the extraction chamber, possibly due to the relatively cool thermal 201
extraction temperature (ca. 100 ºC) and/or cold spots in the transfer line. 202
Improvements to the inlet system may overcome these problems, and thus 203
enhance analytical sensitivity and the recovery of high–molecular-weight 204
recovery. While increasing the thermal extraction temperature may improve 205
extraction efficiency, it may lead to unintentional thermal decrepitating of 206
some of the inclusions. Coupling fluid inclusion extraction with time of flight 207
– secondary ion mass spectrometry for the detection of biomarkers as 208
proposed by Siljeström et al. (2009) may be another way of significantly 209
enhancing the sensitivity of the system, although this approach will lack the 210
chromatographic resolution currently used to obtain a maximum of 211
Volk et al single inclusion analysis_revised_Not Tracked.doc 20/09/2010 page 10 of 16
information on petroleum. Nevertheless, this pilot study proves the concept 212
that ultrashort laser pulses can be used to liberate unaltered oil from 213
individual fluid inclusions, which opens exciting research opportunities to 214
de-convolute information on the nature of different hydrocarbon palaeo-215
fluids from single minerals. 216
217
Acknowledgments 218
We want to thank Neil Sherwood and Roger Boudou (CSIRO Petroleum, 219
Sydney) for their assistance in microscopy and photomicrography and V. 220
Suchý (Czech Academy of Science) for sample material. The paper benefited 221
significantly from comments of journal reviewers Paul Greenwood 222
(University of Western Australia) and Jonathan Watson (Open University, 223
UK), and Guest Editor David McKirdy . 224
225
References 226 Burruss, R.C., 1981. Hydrocarbon fluid inclusions in studies of sedimentary 227
diagenesis. In: L.S. Hollister and M.L. Crawford (Eds) Fluid 228
inclusions: applications to petrology, Mineralogical Association of 229
Canada Short Course Notes 6, pp. 138-156. 230
Fuerbach, A., Fernandez, A., Apolonski, A., Fuji, T., Krausz, F., 2005. 231
Chirped-pulse oscillators for the generation of high-energy 232
femtosecond laser pulses. Laser and Particle Beams 23, 113-116. 233
Volk et al single inclusion analysis_revised_Not Tracked.doc 20/09/2010 page 11 of 16
Gattass, R.R., Mazur E., 2008. Femtosecond laser micromachining in 234
transparent materials. Nature Photonics 2, 219-225 235
George, S.C., Volk, H., Ahmed, M., 2007. Geochemical analysis techniques 236
and geological applications of oil-bearing fluid inclusions, with some 237
Australian case studies. Journal of Petroleum Science and 238
Engineering 57, 119-138. 239
Greenwood, P.F., George, S.C., Wilson, M.A., Hall, K., 1996. A new 240
apparatus for laser micropyrolysis gas chromatography mass 241
spectrometry. Journal of Analytical and Applied Pyrolysis 38, 101-242
118. 243
Greenwood, P.F., George, S.C., Hall, K., 1998. Applications of laser 244
micropyrolysis - gas chromatography -mass spectrometry. Organic 245
Geochemistry 29, 1075-1089. 246
Hode, T., Zebuhr, Y., Broman, C., 2006. Towards biomarker analysis of 247
hydrocarbons trapped in individual fluid inclusions: First extraction 248
by ErYAG laser. Planetary and Space Science 54, 1575-1583. 249
Keller, U., 2003. Recent developments in compact ultrafast lasers. Nature 250
424, 831-838. 251
Lenzner, M., 1999. Femtosecond laser-induced damage of dielectrica. 252
International Journal of Modern Physics B 13, 1559-1578. 253
Volk et al single inclusion analysis_revised_Not Tracked.doc 20/09/2010 page 12 of 16
Martin S, Hertwig A, Lenzner M,: Krüger, J., Kautek, W., 2003. Spot-size 254
dependence of the ablation threshold in dielectrics for femtosecond 255
laser pulses. Applied Physics A – Material Science & Processing 77, 256
883-884 2003 257
Munz, I.A., 2001. Petroleum inclusions in sedimentary basins: systematics, 258
analytical methods and applications. Lithos 55, 195-212. 259
Schubert, F., Diamond, L.W., Tóth, T.M., 2007. Fluid-inclusion evidence of 260
petroleum migration through a buried metamorphic dome in the 261
Pannonian Basin, Hungary. Chemical Geology 244, 357-381. 262
Siljeström, S., Hode, T., Lausmaa, J., Sjövall, P., Toporski, J., Thiel, V., 263
2009. Detection of organic biomarkers in crude oils using ToF-SIMS. 264
Organic Geochemistry 40, 135-143. 265
Thiéry, R., Pironon, J., Walgenwitz, F., and Montel, F., 2000. PIT 266
(Petroleum Inclusion Thermodynamic): a new modeling tool for the 267
characterization of hydrocarbon fluid inclusions from volumetric and 268
microthermometric measurements. Journal of Geochemical 269
Exploration 69, 701-704. 270
Volk, H., 2000. Source rocks, bitumens and fluid inclusions from the Prague 271
Basin (Barrandian, Czech Republic) - Constraints for petroleum 272
generation and migration from petrology, organic geochemistry and 273
basin modelling. Berichte des Forschungszentrums Jülich 3825, 1-274
Volk et al single inclusion analysis_revised_Not Tracked.doc 20/09/2010 page 13 of 16
337. 275
Volk, H., Horsfield, B., Mann, U., Suchý, V., 2002. Variability of petroleum 276
inclusions in vein, fossil and vug cements - a geochemical study in the 277
Barrandian Basin (Lower Palaeozoic, Czech Republic). Organic 278
Geochemistry 33, 1319-1341. 279
280
Volk et al single inclusion analysis_revised_Not Tracked.doc 20/09/2010 page 14 of 16
Figure captions 280 Figure 1: Cross-plot of dependency of work per pulse and pulse frequency. 281
The use of a Herriott-type multipass cell leads to an increase of the 282
cavity length of a femtosecond oscillator and hence the energy per 283
pulse. 284
Figure 2: Petroleum trapped in fluid inclusions in quartz sample KQ7-Qz 285
before (a) and after (b) ultrafast laser ablation. UV epifluorescence 286
photomicrographs. (c) Extraction crater ablated into quartz, scanning 287
electron microscope photomicrograph. (d) Total ion chromatogram 288
(TIC) of the extracted petroleum. nC10 = n-decane etc., iC15 = 289
isopentadecane etc., Pr = pristane). (e) TIC and selected m/z 290
chromatograms providing more detail for compounds in the C5-C9 291
elution range. MCP = methylcyclopentane, B = benzene, MH = 292
methylhexane, DMCP = dimethylcyclopentane, MCH= 293
methylcyclohexane, T= toluene, Alk-CH = alkylcyclohexanes, EB = 294
ethylbenzene, m+p+oX = meta-, para- and orthoxylene.295
Volk et al single inclusion analysis_revised_Not Tracked.doc 20/09/2010 page 15 of 16
296
Figure 1 297 298
299
300
Volk et al single inclusion analysis_revised_Not Tracked.doc 20/09/2010 page 16 of 16
300
Figure 2 301
302
303