The Molecular Weight and Molecular Structure of Asphaltenes.

VPO There are many methods that have been employed to derive the molecular weight of asphaltenes; this diversity is one of the principal reasons for reinforcing strong discord in the literature. One of the most common techniques that have been used is VPO (vapor pressure osmometry). This technique attempts to measures the number of ‘units’ of nonvolatile impurity in a solvent. Asphaltenes undergo aggregation at the concentration that VPO requires so the unit that VPO measures is an aggregate weight. Consequently, the ‘molecular weight’ obtained from VPO depends on solvent, concentration and temperature. Since real molecular weight depends on none of these factors, VPO is not a direct method to obtain molecular weight. Other factors complicate VPO including the possibility that asphaltenes exhibit different levels of aggregation. This has been suggested by comparison of ultrasonic data and other techniques including surface tension data. If so, it might be that extrapolation of VPO data to zero concentration (within concentration ranges where VPO is operative) does not asymptote to molecular dispersion. VPO also has an issue of activity vs concentration. From my point of view VPO should not be used to obtain the molecular weight of asphaltenes, unless it can operate at ~ 100 mg/liter concentration [1]. From my point of view, it is time to stop claiming that VPO gives molecular weight of asphaltenes; it does not.

GPC Gel Permeation chromatography has two debilitating problems. The standards used are typically polystyrene which is a very different material compared to asphaltenes. In addition, the concentrations used may exceed the aggregation threshold as determined by the ultrasonics.[1] From my point of view GPC is not reliable for determination of asphaltene molecular weight.

Optics Techniques. Optical range of Interest for Asphaltenes: Asphaltenes are highly absorptive in the visible even into the near-infrared, while standard polycyclic aromatic hydrocarbons with four fused rings or fewer are all colorless. This tells us that asphaltenes certainly possess some large chromophores. We have employed optical techniques to assess roughly asphaltene molecular weight. The first question is over how broad a spectral range do we need to consider. Optical absorption experiments place limits on the long wavelength end. If there is no absorption at beyond certain long wavelengths by asphaltenes, then this wavelength is of no concern for asphaltenes. We have established that the asphaltene electronic absorption edge (on the long wavelength side) is characterized by the ‘Urbach tail’ in the Fermi edge - a result familiar from solid state physics. Essentially, the electronic edge of various materials exhibits thermal excitation so the electronic edge has an exponential decline with slope kT on the long wavelength side. All crude oils [2] and all asphaltenes [3] have very similar electronic edge slopes characterized by ~10 kT. The electronic edge of these materials is determined by the thermal production of big chromophores from small chromophores, and bigger chromophores absorb at longer wavelengths. The population of big chromophores exponentially declines in asphaltenes. This exponential decline starts around 650nm which roughly (from fluorescence depolarization is ~1000 amu). This result is in agreement with mass spec studies [4] in terms of rapid decline of population in asphaltenes above ~1000 amu.

On the short wavelength side, we cannot look for optical absorption to define the smallest chromophores because big chromophores absorb at short wavelengths, but we can look for fluorescence emission because only small chromophores emit short wavelength light. All fluorescence spectra of asphaltenes, lack much emission from aromatics with one and two rings, at 290 nm and 310 nm respectively. Asphaltenes also lack much fluorescence emission from three-ring aromatics, but there is more emission here than for one- and two-ring aromatics. This is widely known and repeated in all laboratories that measure fluorescence spectra of asphaltenes. In our lab, we probed the reason for this lack of emission from small fused ring systems. Either they are not present in abundance in asphaltenes or they are present but predominantly undergo radiationless emission in asphaltenes. If the former explanation is correct, the lifetime of the UV fluorescence of asphaltenes should match that of maltenes. If the latter explanation is correct, then the small molecules have a new decay path (radiationless transition) thereby shortening the fluorescence lifetime as we have shown [5]. We found that essentially that the UV emission from the asphaltenes and maltenes are comparable, thus we conclude that asphaltenes lack UV fluorescence emission because for the most part they lack one-, two- and largely three- fused ring aromatics [6]. In any event, now we know the relevant spectra range to interrogate asphaltenes; 400 nm to 650 nm.

Fluorescence Depolarization. We employ fluorescence depolarization to determine rotational correlation measurements. This will allow us to determine the size of the asphaltene molecules. First we measure the rotational correlation time tc of two standard molecules, octaethyl porphyrin (OEP) and solar dye (SD) a seven ring aromatic with long alkane chains. For OEP we obtain tc equal to the published value from a very different technique - gamma ray-gamma ray perturbed angular correlation [7]. We then measure a series of asphaltenes, resins, asphaltene solubility fractions, etc [8-13]. We show that asphaltene correlation times are very small - comparable to SD and OEP [8-13]. The mean molecular weight we obtain for virgin crude oil asphaltenes is ~750 g/mole. We show that resin molecules with the same spectral properties exhibit shorter correlation times but comparable to asphaltenes; that is, they are not an order of magnitude smaller, again at the same wavelength of interrogation.

We show that small asphaltene chromophores rotate up to 10 times faster than large asphaltene chromophores [8-13]. This means they are not attached in any cross linked structure. We show that coal asphaltenes are much smaller than petroleum asphaltenes [10,13] and we show as expected (at least some) coal asphaltenes lack alkane chains. We confirm that coal asphaltene aromatic ring systems are smaller than those of petroleum asphaltenes by direct imaging of these asphaltenes using High-Resolution Transmission Electron microscopy [14]. If asphaltenes were cross-linked, then there would be no reason why coal asphaltenes are necessarily much smaller. In addition, coal asphaltenes clearly have smaller ring systems as shown by fluorescence spectroscopy and fluorescence depolarization. The explanation makes sense only if there is a single fused ring system per molecule. We explain this observation based on solubility. We view asphaltene solubility to be governed by the balance of intermolecular attraction vs alkane steric disruption. The attractive forces are dominated by the van der Waals interaction of the different p systems, the bigger the ring systems the stronger the intermolecular attractive forces. Because coals lack alkanes, then they must have small ring systems as we observe to maintain the balance.

In addition, the smallest n-heptane virgin petroleum asphaltene we have ever measured corresponded to an asphaltene that uniquely (for our asphaltenes) showed a very large sulfoxide content. This asphaltene had 44% of its sulfur as sulfoxide, while all of our other asphaltenes have sulfoxide below 5% [15]. The sulfoxide asphaltene is similar to a bidentate ligand. The sulfoxide (which we know is alkyl sulfoxide) is a binding site and the single fused ring system in the molecule is a binding site. Since the sulfoxide has a very large dipole moment (~4 debye), then the ring system must be smaller to maintain constant binding. This makes no sense if there are a large number of ring systems in the molecule since intermolecular binding would be governed by the size and number of fused ring systems.

For FD studies, it is important to note that NMR studies show that average alkane chain lengths are on the order of 4 to 6 carbons [16], it is unlikely that large groups in close proximity do not impact rotational diffusion, especially if cross linked.

Consequently, for both coal vs petroleum asphaltenes and for the sulfoxide asphaltene, the observations are explainable only if there is (essentially) a single fused ring system in each molecule. I point out, the occasional molecule with an alkyl bridge to a small ring system is not prohibited by any of these observations. However, fluorescence spectroscopy, FD, STM, 13C NMR and HRTEM are all consistent with the observation that asphaltene ring systems on average contain 7 fused rings. (see below)

Asphaltene Fused Ring Systems. The FD results show that the correlation times for the asphaltene fused ring systems are comparable to molecules with 4 to 10 rings. The aromatic ring systems in asphaltenes are pericyclic [17] in accord with expectations based on chemical stability [18]. This optical absorption and emission spectra of asphaltenes corresponds roughly to 4 to 10 rings pericyclic rings [18]. This finding is in agreement with direct molecular imaging of asphaltenes by Scanning Tunneling Microscopy [19,20]. In addition, direct imaging of these asphaltenes using High-Resolution Transmission Electron microscopy [14].

C13 NMR studies have been performed on a variety of n-heptane petroleum asphaltenes. These different studies differentiate peripheral and internal carbon in aromatic ring systems. Direct comparison of asphaltenes with a variety of PAH’s shows that asphaltenes have an average of about 7 fused rings in the aromatic core [16]. There is of course, considerable width to this distribution, often quoted as 4 to 10 rings. The Florida State mass spectral group cannot resolve structure at this time but they can determine the number of aromatic rings in the asphaltenes. The find a range from 2 to 12 aromatic rings in asphaltene molecules [4].

Mass Spectroscopy Mass spectroscopy is perhaps the most obvious candidate to determine asphaltene molecular weight. Boduszynski published results field-ionization mass spectroscopy (FIMS) on n-heptane asphaltenes reporting a mean asphaltene molecular weight of ~700 g/mole [21]. These results were at odds with conventional wisdom and so were questioned based on two issues, the ability to obtain gas phase of large components and possible fragmentation. Laser desorption mass spectroscopy (LDMS) and matrix-assisted laser desorption ionization (MALDI) studies were subsequently utilized to study asphaltenes. Both of these techniques are complicated by severe baseline issues. Reports in the literature vary by more than a factor of 10 on asphaltene molecular weight. Some studies obtained values quite close to those of Boduszynsky [21], others much higher. Treatment of the baseline, what is real vs artifact, is perhaps the primary issue separating these different studies. One LDMS manufacturer was shown an LDMS spectrum of asphaltene that seemed to show a large molecular weight. The manufacturer of the equipment, Bruker, instructed that the signal was “noise” and not real [4].

Other ionization techniques have been employed to determine asphaltene molecular weight. Fortunately, these techniques are in agreement with the original FIMS results and with the FD results. These other ionization methods include ESI-FT-ICR and APCI. It is important to note the ESI (the inventor of which won the Nobel Prize) does not ‘evaporate’ the asphaltene. Rather the solvent is evaporated leaving the original solution with no more solvent, thus the solute in a vacuum. In addition, the method of ionization is very soft; there is no fragmentation. Very delicate and heavy systems can be successfully studied with ESI. Asphaltenes are not pushing this technique to the limits. ESI is applicable only on heteroatom containing molecules. The FSU work on asphaltenes is not published yet, but to date they are finding results with ESI very consistent with FIMS and FD [4]. APCI applied to asphaltenes in different laboratories are producing consistent results again [24,25]. And it is important to note that some MALDI studies are consistent with these studies.

A question arises why ESI should issue the same molecular weight as nonheteroatom containing asphaltenes, that there is Bob Long’s inverse relation between polarity and molecular weight. In fact, the sulfoxide asphaltene above proves Long’s point. However, thiophene is not very polar. Even sulfides are not very polar (consider boiling points of mercaptans). Heteroatom containing does not necessarily mean very polar.

The molecular structure results from 13C NMR, STM, HRTEM and Fluorescence spectroscopy are in direct agreement with the molecular weight determinations from Fluorescence Depolarization and mass spectroscopy techniques FIMS, ESI-FT-ICR, APCI and some MALDI measurements. If one constructs a pericyclic fused ring systems with 7 rings, with 40% aromatic carbon, 60% saturated carbon, with one S and one N atom, the resulting molecular weight is ~ 750 g/mole. This is good news!

Solubility. I point out that this issue needs to be considered for asphaltenes. If one accepts the 13C NMR data, the optical data, and the direct imaging methods of STM and HRTEM, then 7 fused rings is the mean for petroleum asphaltenes. Coronene and other large PAH’s have very little solubility in toluene. As mentioned earlier, the alkane chains of asphaltene help dissolve this material. Also, as mentioned earlier, coal asphaltenes do not possess much alkyl carbon so they have smaller ring systems. If 7 fused rings is correct for petroleum asphaltenes, then solubility in toluene will become a problem for more than about one fused ring system per molecule.

Degradation Studies. I will say very little here, I am sure others will be happy to say much. I do point out some pyrolysis studies of athabasca asphaltenes at 400^oC without hydrogen show rapid coking of >50% of the carbon, significant gas production, especially of heteroatom containing molecules, along with resin and oil production [26]. It seems to me these results are inconsistent with the single core picture of asphaltenes. Coking half the carbon where the asphaltenes are about half aromatic carbon is consistent with large condensed cores. Resins could be formed for instance, upon losing a peripheral heteroatom functionality, one could imagine the remaining molecule becoming a resin. The point was made repeatedly in Banff that at least some degradation studies did not speak to the asphaltene molecular weight [26]. And it is highly desirable that a single molecular framework be compatible with all data above and the reactivity data. I would make the case that it is easier to predict hamburger by understanding a cow rather than the other way around, so I would propose to start with asphaltene model compounds that satisfy the above structural data and then develop modifications of those models required by reactivity data. Perhaps we get a little lucky and this process converges.

References:
[1] Ultrasonic Determination of the critical micelle concentration of asphaltenes and other surfactants, G. Andreatta, N. Bostrom, OC Mullins, submitted Langmuir (the title will change from micelle to nano-aggregate for asphaltenes)
[2] Electronic absorption edge of petroleum, O.C. Mullins, S. Mitra-Kirtley, Y. Zhu, Appl. Spectros. 46, 1405 (1992)
[3] First observation of the Urbach tail in a multicomponent organic system, O.C. Mullins, Y. Zhu, Appl. Spectros., 46, 354 (1992)
[4] Ryan Rodgers, Private communication
[5] Fluorescence lifetime studies of crude oils, Xin Wang, Oliver C. Mullins, Applied Spectrosc., 48, 977 (1994)
[6] Small population of one to three fused-ring aromatic molecules in asphaltenes, C.Y. Ralston, S. Mitra-Kirtley, O.C. Mullins, Energy & Fuels, 10, 623 (1996)
[7] Perturbed angular correlation studies of indium metalloporphyrin complexes, Oliver C. Mullins and Morton Kaplan, J. Chem. Phys. 79, 4475 (1983)
[8] Asphaltene molecular size and structure, Henning Groenzin, O.C. Mullins, J. Phys. Chem. A, 103, 11237, (1999)
[9] Molecular sizes of asphaltenes from different origin, H. Groenzin, O.C. Mullins, Energy & Fuels, 14, 677 (2000)
[10] The Overriding chemical principles that define asphaltenes, E. Buenrostro-Gonzalez, H. Groenzin, C. Lira-Galeana, O.C. Mullins, Energy&Fuels, 15, 972, (2001)
[11] Molecular size Asphaltene molecular size for solubility subfractions obtained by fluorescence depolarization, H. Groenzin, O.C. Mullins, S. Eser, J. Mathews, M.-G. Yang, D. Jones, 17, 498, Energy&Fuel, (2003)
[12] Effect of hydrotreatment on asphaltene fractions, L. Buch, H. Groenzin, E. Buenrostro-Gonzalez, S.I. Andersen, C. Lira-Galeana, O.C. Mullins, Fuel, 82, 1075, (2003)
[13] Comparison of coal and petroleum asphaltenes, S. Badre, C.C. Goncalves, OC Mullins, submitted, FUEL
[14] High-resolution transmission electron microscopy of asphaltenes; alkane induced disorder, A. Sharma, H. Groenzin, A. Tomita, O.C. Mullins, Energy&Fuel, 16, 490, (2002)
[15] Determination of the chemical environment of sulfur in petroleum asphaltenes by X-ray absorption spectroscopy, Geoffrey S. Waldo, Oliver C. Mullins, James E. Penner-Hahn, Stephen P. Cramer, Fuel, 71, 53 (1992)
[16] Structural characterization of asphaltenes of different origin, V. Calemma, P. Iwanski, M. Nali, R. Scotti, L. Montanari, Energy & Fuels, 9, 225, (1995)
[17] Carbon K-edge X-ray Raman spectroscopy supports simple yet powerful description of aromatic hydrocarbons and asphaltenes, U. Bergmann, H. Groenzin, O.C. Mullins, P. Glatzel, J. Fetzer, S.P. Cramer, Chem Phys. Lett. 369, 184, (2003)
[18] Y. Ruiz-Morales, J. Chem. Phys. A 106, 11283, (2002)
[19] Molecular Imaging of petroleum asphaltenes by STM: verification of structure from 13C NMR and 1H NMR data, G.W. Zajac, N.K. Sethi, J.T. Joseph, Scan. Micros. 8, 463, (1994)
[20] N. Battina, Mexico conference (2003)
[21] M.M. Boduszynski, Ch. 7, ref. 4. in J.W. Bunger, N.C. Li, Eds., “Chemistry of Asphaltenes,” American Chemical Society, Washington D.C., (1984)
[22] J.T. Miller, R.B. Fisher, P. Thiyagarajan, R.E. Winans, J.E. Hunt, Energy & Fuels, 12, 1290, (1998)
[23] M.-G. Yang, S. Eser, ACS Reprints, ACS New Orleans Meeting, 768, (1999)
[24] B. Desmazières, I. Merdrignac, O. Laprévote, P. Terrier, The Analysis of Oil Asphaltenes by SEC and SEC-MS Coupling, Banff Conference 2004
[25] R.I. Cunico, E.Y. Sheu, O.C. Mullins, Molecular weight measurement of UG8 asphaltene by APCI mass spectroscopy, Petroleum Science and Technology, in press Vol. 22, Numbers 7 & 8, (2004)
[26] K. Soodhoo, C.R. Phillips, Non-catalytic hydrocracking of asphaltenes, Fuel, 67, 361, (1988)

Oliver Mullins
Fluids Product Champion
Schlumberger
Houston, TX