This article was first published in German as "First Direct Chemical Analysis of Interstellar Dust" in Sterne und Weltraum p 326-329, v 39, May 2000 (1). This English translation is published here by permission of coauthor Dr. Jochen Kissel. He comments, "we see this as a way how interstellar dust and comets may have seeded life on earth: certainly not by bringing highly sofisticated material like RNA or DNA directly to earth, but rather by providing the physical and chemical environment for it to develop." Dr. Kissel suggested the title used here. We have lightly edited the translation where it seemed awkward.

The Physical and Chemical Properties of Interstellar Dust and Dust in Comets: Possible Seeds for Life on Earth

by Franz R. Krueger * and Jochen Kissel +

+ Max-Planck-Institut fur Extraterrestrische Physik, Giessenbachstrasse, D-85740 Garching / Germany
* Ingenieurbureau Dr. Krueger, Messelerstr.24, D-64291 Darmstadt / Germany

Jochen Kissel holds CIDA


The dust impact mass spectrometer CIDA onboard the NASA spacecraft STARDUST has detected five interstellar particles and recorded mass spectra during its first measuring period. Due to their unexpected complexity the analysis is an arduous task and requires new methods. Nevertheless, the dominating substance class — namely, polymeric hetercyclic aromates — and the particle masses could be determined. Both together are consistent with optical properties of those interstellar particles being able to reach the inner solar system.


On the 7th of February 1999 the spacecraft STARDUST was launched from Cape Canaveral/FL, after a one-day delay, to its long journey to comet p/Wild-2, and subsequently back to earth. Its main task in January 2004 is sampling cometary dust by aerogel plates made from silicon dioxide of very low density (approx. 0.03 g/ml), bringing them back to earth in January 2006. However, it is expected that the organic component of the dust will be destroyed during collection by the aerogel. As a result the information gathered will be limited to the mineral component, and to the isotopic distribution of all elements contained in the dust, except for silicon and oxygen, of course.

In order to perform an in-situ chemical analysis of the cometary dust during encounter we implemented a dust particle impact mass spectrometer onboard, named CIDA (Cometary Impact Dust Analyzer). Not only will CIDA analyze the molecular composition of the dust of comet p/Wild-2, but also that of the interstellar dust intercepted during its flight from earth to the comet and back to earth again. Due to the fact that the trajectories of the interstellar dust streams are often perpendicular or at least inclined relative to those of the interplanetary dust stream, there are periods of some months each during which the instrument is sensitive to interstellar dust exclusively. One of those periods is already completed, and the results are reported here. Before encounter with p/Wild-2 there are two more such periods, and, if all runs well, others will occur during the spacecraft's travel back home.

The measuring principle of CIDA

Already as early as 1986 three of our dust impact mass spectrometers (PIA on board GIOTTO, and PUMA-1 and -2 onboard VEGA-1 and -2) encountered comet p/Halley. With those we were able to determine the elementary composition of the cometary dust. Molecular information, however, was only marginal. This was due to the fact that the relative velocities between each spacecraft and the retrograde comet were extremely high (GIOTTO: 69 km/s; VEGA: 78 km/s). Because p/Wild-2 is a prograde comet, the velocity relative to STARDUST is only about 6 km/s. The relative velocities of interstellar streams are in the order of 25 km/s. In these velocity ranges CIDA is more sensitive to molecular rather than atomic information.

Depending on direction and position of the CIDA instrument, interstellar, interplanetary, or cometary particles may impact on the about 100 sq-cm circular silver target (Fig.1). During impact they instantaneously evaporate forming molecular and atomic fragments (depending on impact speed, as already mentioned), which are in part electrically charged, either positively or negatively. By a voltage of U=1 kV applied between target and acceleration grid either positive or negative ions are accelerated. Due to the fact that the main chemical information is "contained" in the positive ions, only those will be measured until encounter with the comet. Moreover, changing polarity for detection of negative ions is risky due to the higher voltage of the instrument relative to the ion detector (which in any case measures secondary electrons after cathodic conversion). After encounter perhaps also negative ions can be measured during some flight periods. As a result special information about the oxygen- and sulfur-chemistry of some organic functional groups (or mineral parts) may be gained.

After acceleration all ions (being exclusively singly charged with the elementary charge e) of different molecular (or atomic) masses m (in daltons) possess the same kinetic energy E = 1/2 m v^2 = eU . Consequently, each ion's velocity v in the subsequent drift space is inversely proportional to the square root of its mass. However, a certain velocity distribution width is due to different initial energies of the ions produced. This uncertainty is compensated in first order by an electrostatic reflector — ions traveling "too fast" due to their higher initial energy will fly deeper into the reflector than others of the same mass. This longer flight path causes them to arrive at the detector at the same time as the slower ones of same mass. Thus the total travel time (from the target all the way to the detector) t of each ion of mass m is directly proportional to the square root of m. This yields t = a*sqr(m) + b; with "a" being a function of the flight path effective length and the voltages applied; in principle, "b" is given by the electronic signal travel times relative to the time t=0 for (an imaginary) mass m=0. The time spectrum of the ion current at the detector can thus be converted into a mass spectrum, exact knowledge of a and b provided. (If the impact time is unknown, there are great mathematical difficulties to overcome in order to determine b.) The mass resolution m/delta-m is about 300. Thus the mass line of m=300 is still fairly resolved from that of m=301. The quality of the above transformation "time into mass" is best if there are only spectral peaks due to INTEGER mass numbers AND a and b are in agreement with the actual parameters of the instrument. The chemical information about the dust particle impacted is then "hidden" in the intensity distribution of the detector signals over the entire mass range.

The analysis of the mass spectra

When analyzing substance mixtures in the laboratory, generally those mixtures are first (e.g., chromatographically) separated, and subsequently each single substance is analyzed mass spectrometrically. Naturally, any pre-separation is impossible with such swift dust particles. The mass spectrum generated by a particle impacting on the CIDA target thus represents the mixture within the particle as a whole. Consequently, one can only hope that at least an analysis of substance CLASSES comprising the majority of the particle's matter is possible. A necessary condition for class analysis is that the mixture not be too heterogeneous chemically. However, even in a favorable case this class type analysis presents a completely new challenge. [In this very context it is helpful that we and a group of Viennese chemists are just developing a chemometric method for the purpose of analyzing functional groups in substance mixtures. As early as 1987, we made use of a conceptually similar, simpler method in analyzing the molecular ion contribution in p/Halley's dust impact mass spectra. However, a further methodological development is needed for our experiment, COSIMA (COmetary Secondary Ion Mass Analyzer), onboard the ESA spacecraft ROSETTA to comet Wirtanen, which will be launched in 2003. In rendezvous with that comet we plan to collect cometary dust on metal-black targets and subsequently analyze it by Secondary Ion Mass Spectrometry (SIMS). As we know from laboratory experiments, the chemical process of ion formation in SIMS is quite comparable to that in dust impact in the lower velocity (below 25 km/s) range.]

The chemometric method of analysis of substance classes is actually a type of autocorrelation analysis. Take as an example the alkanes that are abundant in fuels, like propane, butane, pentane, hexane, octane — and paraffins, which are higher alkanes. All linear alkanes are expressed by the general chemical formula Cn H(2n+2) (n=1,2,3,...). With the mechanisms of ion formation from the condensed phase (SIMS, Impact Ioniz., etc.) they frequently form non-radical (i.e. even-electronic) ions of the common form Cn H(2n+1)+ directly as a quasi-molecular ion. However, decomposition by elimination of CH2-groups leads to the same type of ions. A mass spectrum of such a mixtue of alkanes is characterized by a strong abundance of mass lines corresponding to ions whose molecular masses differ by a fixed delta-m = 14. The mass line basis is m=15, i.e., n=1 (CH3+). Pure carbon, as a further example, would show up with characteristic mass differences delta-m = 12 with the mass basis 12: C+, C2+, C3+, .... An over-statistical representation of delta-m = 48 (difference of 4 C- atoms) indicates graphite or large PAH's (Polyaromatic hydrocarbons) with little contribution of hydrogen in ion formation. Diamonds ionize with an expressive even-odd pattern of the numbers of carbon atoms in the ion; thus in this case, delta-m = 24 is more representative than delta-m = 12 is. If, however, delta-m = 24, 48 is not statistically overrepresented in comparison to delta-m = 12, this would indicate amorphous carbon, which, by the way, does not tend to form large molecular ions, in contrast to the other modifications of carbon. Having said this, we did not find among our interstellar particles indications for these substance classes just mentioned.

There are additional chemical classifying methods other than modulo-functions found by autocorrelations; however, these are not being treated here.

The dominant substance class of the interstellar particles

An important question is whether silicates may form the dominant substance class. Silicates mainly contain the elements O, Mg, Si, Ca, and Fe. Other minerals may contain C (carbides, carbonates) or S (sulfides), additionally or alternatively. Well, we know how those elements are cooked in certain stars; their most abundant isotopes are mainly aggregates of He-4 nuclei. Even if not all, their integer atomic mass numbers are factors of 4 anyway. Thus the mass numbers m of atomic and molecular ions thereof are also multipless of 4. Consequently, in this case we expect their mass spectra to be strongly autocorrelated with delta-m = 4 with a basis mass number dividable by 4, too. This is found in none of the mass spectra! Silicon dioxide, silicates, as well as iron ores, can thus be excluded as dominant components of our interstellar particles. Examining the first two impact events was a great surprise. Mass numbers up to and beyond m=1000 ions were present, even quite abundant. From the total ion number and its distribution over m we were able to estimate the total mass of the impacting particles. The estimated masses of the two first particles were somewhat below 1 nanogram each. Furthermore, the magnitude of their ion masses strongly suggests that the ions come from a highly polymerized matter, perhaps similar to plastics which are built from various copolymers. The next three particles to impact the instrument were apparently lighter, about 10 picograms each. Their ion masses were distributed up to about m=400. However, this does not mean that they are less polymerized. Rather, their lower total mass caused the detector ion current to be lower, just dropping under threshold in the higher ion mass range (fig.2). A further surprise was that we did not see many impacts of smaller particles, although CIDA's detection threshold is well below 100 femtograms. Usual particle size distributions show more abundance at lower masses! The solution to this enigma came from totally different arguments — not mass spectrometry — and will be discussed in the last section.

What are the functional groups of the giant molecules we found in the interstellar dust, and how did we ascertain them? After having excluded the dominance of several elements, we were left with only 5 elements which could have built up the majority of these polymers, namely: Hydrogen (H), carbon (C), nitrogen (N), oxygen (O), and (unsure) sulfur (S). The integer mass numbers of stable molecules built from these elements except nitrogen (!) are even! (This is due to the even valence numbers of C, O, and S and the even mass number of their main isotopes.) Abundant non-radical ions (positively charged by protonization, negatively charged by deprotonization) thus possess odd mass numbers! Both laws are still true, if the molecule or ion contains an even number of N atoms, as can be easily verified. Thus, delta-m = 2 is the strongest auto-correlation, as being well established in all 5 impact mass spectra. The main basis mass numbers are thus odd, especially with the 3rd and 5th impact: with these particles nitrogen apparently plays a minor role only. (see fig.3a) The subsequently found correlations and basis masses there indicated an unsaturated-carbon chemistry with some oxygen contribution. Furanes or dioxines may be subgroups of those polymers; their main constituents, however, are more or less reduced (hydrogenized) homocyclic aromates which may be also bound together via oxygen bridge bonds (ether-type binding). It is important to note here that those (only in part aromatic) polymers are steric. I.e., their conformation deviates strongly and erratically from the planar geometry PAH's or graphite. They are probably more similar to tars, rather than crystallized matter.

The contribution of nitrogen, which is low with the 3rd and 5th particle, seems to be larger with the 1st and 2nd particle. This is recognized by the dominance of odd delta-m, esp. 15 and 27 (difference due to an HN- or an HCN-group or eliminant). An elementary structure may result from chinoline or carbazole groups, for instance (fig.3b). But let us now focus on condensed aromates — apparently the main substance class found — with hetero-cyclic components. Our analysis does not point to a homocyclic aromate like coronene (C24H12). An exchange of carbon by oxygen can only take place with the outer ring atoms (or we may have a keto-groups there), thereby breaking locally the mesomeric electronic structure. This molecule would not be planar any more. An inner C-atom cannot be replaced by an O-atom, because it has only two bonds. An N-atom, however, could well replace an outer CH-group isoelectronically without breaking the planar geometry. An inner C-atom, however, can only be replaced by an N-atom, disturbing electronics and conformation, because it causes a steric sp3-hybridic bonding which breaks the mesomeric electronic structure as well. Steric ring annealation is then possible, as we know, for instance, from alkaloids like sparteine, although these are more reduced (i.e., hydrogenated). Dehydrogenated (oxidized), some via N sterically annealed, and via O-bridges bound ring structures are main characteristics of the interstellar dust matter found by CIDA.

Interstellar dust in the solar system

Of course, we do not and we will not have a representative sample of interstellar dust grains impacting CIDA, but get only those which manage to approach our sun down to the 1 - 2 AU distance range. In contrast to cometary dust — which is said to be primarily interstellar — free interstellar dust lacks the cooling by the cometary nucleus. Which free particles we meet depends on their gravitational, electromagnetic, light-optic, and chemical-thermodynamic properties. Particles smaller than some tenths of a micron can hardly absorb sunlight, and thus may travel into the inner solar system without being hindered. However, particles smaller than .2 um do not trigger a CIDA signal, thus we do not see them anyway. A particle larger than a few microns absorbs sunlight more or less, depending on its constitution; some of them (the larger and the more refractive ones) may thus also reach the inner solar system. Medium sized particles (which CIDA would see if they were present) absorbing the visible and near-ultraviolet light from the sun would either be deflected by radiation pressure before reaching the inner solar system, or evaporate (thus becoming smaller). Indeed, Landgraf et al. believe they found such a particle size gap with their dust detectors onboard GALILEO and ULYSSES. In any case, CIDA did not see "small" particles between 10 femtograms and 1 picogram, although it is sensitive in this range. Moreover, if a "normal" dust size distribution were present, the probability of finding many smaller particles in the measuring period would have been high, and smaller particles should be much more abundant than larger ones. Instead, CIDA detected only five larger dust particles. By means of optical calculations (via the complex refraction index), Landgraf was able to show that those particles detected by his instruments could not have been dominated by graphite or silicon dioxide. This is concordant with our results.

Let us repeat — we can only see those fraction of interstellar dust being intercepted by our instrument in the 1 AU distance range from the sun. Less polymer organic substances than those found can survive in grains only in the coldness of far space. They evaporate in the solar system. Graphite grains or those constituted by very large PAH's, if at all present, would be hindered from entering the inner solar system by radiation pressure. Nevertheless, some speculation may be allowed about the formation and constitution of that solid interstellar matter for which entry to our habitat range is prohibited. Looking to the chemistry of original interstellar gas (!) surrounding far stars, a fraction of highly unsaturated carbon chain molecules and the contribution of nitrogen and oxygen to them is remarkable. Hydrogen interacts far less than its abundance there might let us think. These unsaturated molecules, however, are highly reactive, even by simple two-body collisions due to their many inner degrees of freedom. Thus they polymerize and cyclisize in the cooler parts (below 1000 K) of interstellar space. Their aggregation to dust particles, such as those we have seen, is thus easily possible. Also the "diffuse clouds" have optical properties compatible with homologues of such matter.

A lot of investigations are still ongoing. We are lucky to have for the first time matter from alien stars "in our hands," as it may have also formed our solar system billons of years ago. This is the origin of the matter that constitutes life.


CIDA was developed in a very short time, and would not have been ready for launch without the dedication of our contractor von Hoerner & Sulger GmbH in Schwetzingen, Germany, which worked in part under contract with the German Space Research Agency DARA.


3 Dec 2011: The data from every direction support the interstellar life and panspermia hypothesis — Chandra Wickramasinghe
27 Oct 2011: Theoretically, this is impossible, but observationally we can see it happening, says Professor Sun Kwok.
17 Jun 2004: Mass spectrometer on Stardust sees organic matter at comet Wild 2.
9 Jan 2004: Complex organics in interstellar dust confirmed.
3 Jan 2004: Stardust sailed safely by comet Wild 2.
2002, Nov 4: CIDA is not miscalibrated, according Kissel.
CIDA miscalibrated? — the email exchange with Kissel, 4 Nov 2002.
Cometary and Interstellar Dust Analyzer, from the Stardust website, JPL, NASA.
2002, Mar 12: Is CIDA mis-calibrated?
2000, Apr 27: Most interstellar particles captured by Stardust are complex organic compounds, and not the expected minerals — CA's first announcement of the Stardust results.


1. Franz R. Krueger and Jochen Kissel. "First Direct Chemical Analysis of Interstellar Dust," p 326-329 v 39 Sterne und Weltraum May 2000.
COSMIC ANCESTRY | Quick Guide | Next | by Jochen Kissel | All Rights Reserved