The rod translated and rotated by stepper motor laser pulse vaporises rod Helium pulse C 60 reacts with C, C 2 exits and undergo supersonic expansion C 60 coated graphite rod Cluster Source

Helium Pulsed Nozzle Rod coated with 13 C enriched amorphous carbon (ca 10%) and pure C 60 Pulsed Laser

Helium Pulsed Nozzle Pulsed Laser Rotating/translating rod coated with 13 C enriched amorphous carbon (ca 10%) and pure C 60

Graphite rod surface coated with C 60

O OO OO O O O O OO OO O O O O OO OO O O O OO OO O O O OO OO O O O OO OO O O O OO OO O O O OO OO O O O OO OO O O O OO OO O O O OO OO O O O OO OO O O O OO OO O O O OO OO O O O OO OO O O O OO OO O O O OO OO O O O OO OO O O OO OO O O O OO OO O OO OO OO O O O OO OO O O OO OO O O O OO OO O O O OO OO O O O OO OO O O OO OO O O O OO OO O O O OO OO O O O OO OO O O OO OO O O O OO OO O O Carbon atoms only vapourised from the very hot (10000 o ) focal point …C 60 evapourates mainly unscathed from the surrounding warmed domain

accumulation octopole target rod skimmer transfer octopole ICR cell stepper motor pulsed valve torr diffusion pump turbo pump turbo pump turbo pump pulsed valve source

carbon plasma cluster beam Laser and graphite disk skimmer hole to mass spectrometer Laser Vapourisation Cluster Beam System - Smalley

supersonic expansion skimmed into beam C 60 reacts with carbon species

The pulses pass across the chamber at about 10 Hz …and the signal integrated perhaps 100 to a1000 or more times The University of Sussex machine

ions accumulated in octopole 3-10 laser shot accumulated transferred to the ICR cell

The very sharp edged skimmer skims the expanding pulse into a very narrow beam Supersonic Expansion into the vacuum chamber cools the clusters to very low temperatures

Skimmer action more slowly

supersonic expansion skimmed into beam C 60 reacts with carbon species ions accumulated in octopole 3-10 laser shot accumulated transferred to the ICR cell

C 60 coated on a glass rod – just C 60 C 60 Typical desorbed C 60 mass spectrum. No growth

C 60 on graphite rod – little fragmentation small Cn species are ingested by C 60 forming larger fullerenes C 70 Desorbed C 60 Minimal fragmentation of C 60 Significant growth to larger fullerenes

C 60 + amorphous 13 C coated on a quartz rod Desorbed C 60 C 60 ingests 13 C species

Expanded view of larger C 60 + C n fullerenes The larger fullerenes become increasingly 13 C enriched. In agreement with small C n species inserting into C 60 and higher fullerenes to form even higher fullerenes C 62 C 64 C 66 C 68 C 70 C 76 C 74 C 72 C 78 C 80

Expanded view of larger C 60 + C n fullerenes The larger fullerenes become increasingly 13 C enriched. In agreement with small C n species inserting into C 60 and higher fullerenes to form even higher fullerenes C 62 C 64 C 66 C 68 C 70 C 76 C 74 C 72 C 78 C 80 Note C 62 – C 68 are not stable

C 60 + C n → C 62 C 64 C 66 C 68 C 70 C 72 etc This study shows unequivocally that fullerenes can grow by ingestion of smaller carbon species in this case with Paul Dunk and Alan Marshall

Multi stage formation mechanism

Refinement of a closed cage growth mechanism proposed by Heath for the fullerenes and Endo and Kroto for nanotube growth

Multi stage formation mechanism 1.Laser fires and produces a ca10000 o plasma of C atoms

Multi stage formation mechanism 1.Laser fires and produces a ca10000 o plasma of C atoms 2.The atoms at the plasma/He interface cool to form C 2 C 3, C 4 etc

Multi stage formation mechanism 1.Laser fires and produces a ca10000 o plasma of C atoms 2.The atoms at the plasma/He interface cool to form C 2 C 3, C 4 etc 3.Monocylic rings (and another family) form

Multi stage formation mechanism 1.Laser fires and produces a ca10000 o plasma of C atoms 2.The atoms at the plasma/He interface cool to form C 2 C 3, C 4 etc 3.Monocylic rings (and another family) form 4.Small fullerene cages C 28 … are created

Multi stage formation mechanism 1.Laser fires and produces a ca10000 o plasma of C atoms 2.The atoms at the plasma/He interface cool to form C 2 C 3, C 4 etc 3.Monocylic rings (and another family) form 4.Small fullerene cages C 28 … are created 5.Small fullerenes grow into larger cages by ingestion

Multi stage formation mechanism 1.Laser fires and produces a ca10000 o plasma of C atoms 2.The atoms at the plasma/He interface cool to form C 2 C 3, C 4 etc 3.Monocylic rings (and another family) form 4.Small fullerene cages C 28 … are created 5.Small fullerenes grow into larger cages by ingestion 6.Final stage – only C n n=60, 70 and higher n species survive

Graphite rod surface coated with C 60

The pulses pass across the chamber at about 10 Hz …and the signal integrated perhaps 100 to a1000 or more times The University of Sussex machine

The rod translated and rotated by stepper motor laser pulse vaporises rod Valve emits a helium pulse Fullerenes react with carbon vapor (C, C 2 ) in the “clustering zone”. Then, the gas exits the channel and undergoes a supersonic expansion to create a cooled, molecular beam fullerene-coated graphite target rod Cluster Source

Expanded view of desorbed C 60 – normal isotope distribution Normal ratio ← 13 C 12 C 59 ~ 60% of 12 C 60 signal 12 C 60 signal

One strongly held conjecture was that C 60 and C 70 would be cul-de-sacs and our results indicate that this conjecture is incorrect.

Our results indicate that the primary nascent distribution of fullerenes shows almost no evidence of IPR stabilisation which is a surprise to at least me Requiring a final stage which non IPR cages do not survive

One strongly held conjecture was that C 60 and C 70 would be cul-de-sacs and our results indicate that this conjecture is incorrect. Our results indicate that the primary nascent distribution of fullerenes shows almost no evidence of IPR stabilisation

C 60 + amorphous 13 C on a quartz rod Desorbed C 60

Q: How many pulses do you need to accumulate? A: A single laser shot is used vaporize the target during a single Helium pulse. Ten singe laser shot + He pulse are used to accumulate ions. Q:How do you decide when to transfer them. A: After the final laser shot, a voltage at the “back” of the accumulation octopole switched, and the ions are transferred to the ICR cell. The switching of the voltage is controlled by the computer program interface. Q: How many runs do you need in general for an average result? A: 3 time-domain aquisitions are averaged for when “growing” a preformed fullerene…….the signal is extremely strong. And 10 time-domain acquisitions are averaged when form endohedrals from a graphite-metal target. Thus, up to 10 time-domain acquisitions are averaged. Answer to questions from .

Fullerenes react with carbon vapor in the “clustering zone”, then the gas exits the channel and undergoes a supersonic expansion. As the clusters move from a region of high pressure through a small orifice into a high vacuum, they undergo a supersonic expansion. The random thermal energy of the clusters is converted into a directed motion (creating a cooled, molecular beam in which very few collisions occur) toward the skimmer and the ions subsequently enter the ion optics where they are accumulated and then transferred to the ICR cell for detection.

Q: How do you stop the pulse of ions in the accumulation trap A: The ions are confined radially by an oscilating radiofrequency within in octopole, and axially by voltages at the ends of the “accumulation octopole:. Answer to questions from .

…undergo supersonic expansion and are skimmed into beam torr diffusion pump turbo pump turbo pump turbo pump Fullerenes react with carbon species in the vapourisation zone, then exit reaction channel the ions which enter the ion optics, where they are acuumulated in the central octopole segment After 3-10 single laser shot accumulations, the ions are transferred to the ICR cell, which is located within in the bore of a 9.4 tesla superconducting magnet. Under the influence of the high magnetic field, the ions exhibit cyclotron motion. The ions induce a current on electrodes, which is detected as an “image current” in the time domain, and then the signal is converted to the frequency domain by an FT. Thus, the mass of the ion is detected as a frequency.

Expanded view of larger C 60 + C n fullerenes Larger fullerenes increasingly 13 C enriched C 62 C 64 C 66 C 68 C 70 C 76 C 74 C 72 C 78 C 80

2.7 eV C eV C C 2 facile C 30 and higher C 28 C Less reactive C 30 and higher Electron donation stabilizes small fullerenes – prevents addition of small Cn + C 2 Added electron density at the triple pentagon junction stabilizes – prevents C2 addition Empty cage endohedral

Fullerene Growth Mechanism “fullerene road” – Small fullerenes are the first to form, and then growth to larger fullerenes occur by uptake of small carbon species such as C 2. There has been no evidence that fullerene growth can actually occur this way…until now! This growth model accounts for all experimental observations, for endohedrals and empty cages. This is potentially extremely significant if this checks out.

C 60 C 60 has been desorbed many times and analyzed many times by mass spec. No growth to larger fullerenes occur. However, most of these experiments are performed in a vacuum under conditions where fullerene growth will not be significant.

Bulk C 60 coated on a graphite rod The rod was put in our cluster source Importantly, the experiment was performed exactly as I would if I were ablating a “clean” carbon rod to produce fullerenes. Pulse gas, laser timing, etc all known to be the conditions to see fullerenes The result: Addition of C 2 to form larger fullerenes!

Small carbon clusters added to C 60 to form larger fullerenes. The coated C 60 did not significantly fragment, but did significantly add carbon to form larger fullerenes. C 70 Desorbed C 60 Very minor fragmentation of C 60 Significant growth to larger fullerenes

Larger clusters are FULLERENES. C 70 (formed from desorged C 60 ) was SWIFT isolated, and then subjected to collision with He while exciting (SORI). These larger clusters are clearly fullerenes as C 2 fragmentation occurs. C 70 C 68 C 60

C 60 coated on a quartz rod To gain further insight, C 60 was coated on a quartz rod The same experiment was performed to see if fullerene growth occurred. If no growth occurred, the small carbon species from the graphite were likely adding to C 60 in the coated graphite rod experiment. The result: no growth!

C 60 + amorphous 13 C on quartz rod C 60 was mixed was some amorphous 13 C. There was much more C 60 than amorphous 13 C, approximately 3:1. This mixture was applied to a quartz rod(from toluene) Growth to larger fullerene occurred, and they were more 13 C enriched with size. This is consistent with small cluster addition to C 60

n + directly formed from U/graphite target enriched with amorphous 13 C --for a total 13 C content of 10% 28 experimental 28 simulated 28 experimental 28 (9.5% 13 C) simulated C C 36 C 44 BOTTOM-UP GROWTH n + directly formed from U/graphite target

Growth mechanism and endoherals It is shown through our experiments that the classical endohedrals of Sc, La, and now Ti, Hf, Zr, U all strongly form endohedral fullerenes. But fullerenes smaller than C 60 are most abundant, a clear deviation from the empty cages. Ionic model – electrons from the metal are transferred to the carbon cage in endohedral fullerenes giving, essentially, an indissociable salt...the cage is negatively charged, the encapsulating metal is positively charged. Our experiments coupled with the theoretical data show this principle applies to small fullerenes too. Electron transfer from the encapsulating metal to fullerene cage stabilizes the small fullerene from small carbon addition to larger fullerenes. This is why metals that can donate 3-4 electrons to the cage predominately form smaller fullerenes.

Growth mechanism and 28 Our experiments show that 28 forms first. And then larger clusters are seen under conditions that allow more growth. The metal nucleates initial growth. Experiments will need to be performed with coating a rod with an endohedral, I plan to ask Shinohara for a sample to use. This will prove that the “fullerene road” applies to endohedrals as well as the empty cages. Only a tetravalent metal can stabilize C 28 sufficiently to yield a M 4- Our calculations show that the donated electrons reside at the most reactive triple pentagon junction…the end result: C 28 does not completely react to larger fullerenes.