IR‐Laser Ablation of Potassium Cyanide: A Surprisingly Simple Route to Polynitrogen and Polycarbon Species

Abstract Pulsed laser irradiation of solid potassium cyanide (KCN) produces, besides free nitrogen and carbon atoms, the molecular species KN and KC which are potential candidates for interstellar species of potassium. Additionally, N3, N3 −, KN3, C3, C3 −, and KC3 are produced and isolated in solid noble gases as well as in solid N2. Molecular potassium nitrene (KN) reacts with dinitrogen in neon and argon matrices after photochemical excitation (λ=470 nm) forming molecular end‐on (C ∞v) and side‐on (C 2v) potassium azide isomers. The side‐on isomer (C 2v) is thermodynamically favored at the CCSD(T)/ma‐def2‐TZVP level of theory. It can be obtained from the end‐on isomer by UV‐irradiation (λ=273 nm).

Potassium cyanidei so ne of the metal cyanide species detected in proximity to as tar in space [1] and KC, KN, and KO are supposed to be potential candidates for furtheri nterstellar specieso fp otassium. [2] While KC and KO were characterized by their rotational spectra, [3] to our knowledge no spectroscopic gas-phase study has yet been published for KN. However,p otassium nitrene has been predicted to be stable with ab ond dissociation energy of 81 kJ mol À1 at the MRCI(+ Q)/triple-z level of theory. [2] Pulsed IR-laser ablationo fs olid salts was recently shown to be ap articularly suitable method to produce and characterize anionic complexes. Recent examples are the homoatomic anions Cl 3 À and F 3 À ,w hich were produced by co-depositiono f IR-laser ablated potassiumh alide, KX (X= Cl and F, respectively), with gaseous X 2 /noble gas mixtures at cryogenic temperatures. [4] Twof urther homoatomica nions, which are of particular interest for the present work, are the C 3 À and N 3 À anions. The first one was assigned by Szczepanskie tal. to ab and at 1721.8 cm À1 after isolationa nd electron bombardment of laser ablated graphite in as olid argon matrix. [5] The free N 3 À anion was associated with an IR band at 2003.5 cm À1 by Michl and co-workers, which was observed in pure N 2 matrices after atom/ion bombardment. [6] It was later also detected by Andrewsa nd co-workers after co-deposition of laser ablated Ga, In, and Tl atoms in solid nitrogen, [7] and in argon matrices, wheret his band appeareda t1 991.9 cm À1 . [8] As al ogical extension of our former work [4] we present in this study resultso btained by laser ablation of the ternary systemp otassium cyanide (KCN). Matrix-isolation IR-spectra of thermally evaporated sodium and potassium cyanidew ere reportedp reviously by Ismail, Hauge, and Margrave (IHM). [9] The IR spectra obtained by us from IR-laser ablation of potassium cyanide deposited in solida rgon are in the spectralr egion above2 000 cm À1 are very similar to those described by IHM. We find KNCOa nd CO as the main impurities in our spectra due to the high temperature reactiono fK CN with CO 2 formed by laser ablation of CO 3 À impurities. TheC Ns tretching region of the IR spectra of laser ablated KCN isolated in neon and argon is shown in Figures S1 andS 2i nt he Supporting Information. Based on the work by IHM, the observed bands at 2048. 7, 2059.3/2061.3, and 2079.6 cm À1 in argon and 2047.0, 2061.2, and 2075.7 cm À1 in neon are assigned to monomeric, oligomeric, and polymeric potassium cyanide, respectively.I n contrastt ot he thermale vaporation,w hiche ssentially results in isolated monomeric and oligomeric ion pairs, laser ablation allowsf or the preparation of free anions, for example, laser ablation of alkali halides (MX, X = Cl, F) allowed us to study not only isolated MX ionp airs in solid nobleg as matrices, but also to isolate free anions such as the free X 3 À ions in the presence of X 2 . [4] It was, therefore, rather surprising to find that laser ablation of KCN does not produce free CN À ions, which have a   reported band at 2053.1 cm À1 in solid neon [10] (cf. Figure S1, Supporting Information).
In IR spectrao btained from laser ablated potassium cyanide trapped in solid argon at 12 Kw ef ound two new bands in a region at 1722.0 and 1712.5 cm À1 where we do not expect any CN stretching bands ( Figure 2). Both bands reveal al arge 13 C shift of À66 cm À1 in experiments using K 13 CN, and when KCN was replaced by NaCN it became evident that the 1722.0 cm À1 band is metal independent,w hereas the 1712.5 cm À1 band is not. The metal-independent band is close to the band at 1721.8 cm À1 previously assigned by Szczepanski et al. to the antisymmetric stretch (n 3 )o ft he linear C 3 À anion. [5] Here we confirm the assignment of this band based on its metal independency and its 12/13 Ci sotope pattern obtained after pulsed laser deposition of a1 :1 mixture of K 12 CN and K 13 CN in argon ( Figure S6, Supporting information). The 1:1i sotopic mixture yielded six 12/13 C 3 À isotopologues. The band positions of all C 3 À isotopologues are displayed in Ta ble 1.
The metal-dependent band at 1712.5 cm À1 in Figure 2h as almostthe same 13 Cisotopic shift as the free C 3 À anion. It is assigned to the n 3 stretch of the corresponding ion pair KC 3 (C 2v ) which has been predicted to be the most stable potassium doped carbon clusterK C n (n = 1-10) by ar ecent DFT study. [11] Our quantum-chemical calculations at the CCSD(T)/ma-def2-TZVP level also support this assignment (Table 1, Figure 1). The full isotope pattern of KC 3 is not as nicely resolved as for C 3 À due to lower yields ( Figure S6,Supporting Information), so that only three of its isotopologues can be confidently assigned. The band at 1742.3 cm À1 observed after laser ablation of natural NaCN (Figure 2d)i mmediatelys uggests an assignment to NaC 3 since the n 3 (C 3 À )s tretcho fM C 3 is intuitively expected to shift to higherw avenumbers for lighter alkali metals. However, CCSD(T) calculations predict the n 3 (C 3 À )o fN aC 3 16 cm À1 lower than for KC 3 .T herefore, this band must yetr emainu nassigned.
In solid nitrogen matrices no irradiation was needed to produce KN 3 .A fter deposition the 14 Ns pectra showed as trong band at 2048.5 with as ite at 2049.3cm À1 for end-on K 14 N 3 and the broad band of free 14 N 3 À as described by Michl and Andrews at 2003.5 with as ite at 2005.6 cm À1 .U pon annealing to 25 Kt he band at 2048.5 cm À1 increased drastically while the N 3 À band decreaseds ot hat another band at 2006.5 cm À1 becamev isible. The 2006.5 cm À1 band has not been described by Michlo rA ndrews and is assignedt ot he antisymmetric azide stretch (n 3 )o fs ide-on KN 3 in solid nitrogen. In the mixed isotope experiment where laser ablated KC 14 N/KC 15 N( 1:1) was co-deposited with pure 14 N 2 / 15 N 2 (1:1) at 12 K( Figure S7, Supporting Information), the band at 2048.5 cm À1 split into an octet, whereas the band at 2006.5 cm À1 split into as extet,a s expected for end-ona nd side-on KN 3 ( Figure S7b, Supporting Information). Upon irradiation with UV light (l = 273 nm) the free N 3 À band was depleted, while the 2048.5a nd 2006.5 cm À1 bands were not (FigureS7c, Supporting Information). All calculated and observede xperimentalb and positionso ft he antisymmetric N 3 À -stretches of side-on and end-on KN 3 in Ne, Ar, and N 2 are displayed in Ta ble 2.
Calculationsa tt he CCSD(T)/ma-def2-TZVP level of theory predictt hat side-on KN 3 is 3.5 kJ mol À1 lower in energy than end-on KN 3 .T he barrier for the rearrangementf rom end-ont o side-on obtained from ar elaxeds urfaces can along the KÀNÀN bond angle (Figure 4) is predicted to be 12.5 kJ mol À1 which appearst ob eh igh enought hat the rearrangement could not be observed by annealing in any experiments. In neon, however,t here seems to be ap hotochemical equilibrium between the end-on and the side-on form:I mmediately after formation of the two isomersb yi rradiation with blue light (l = 470 nm)  ]o fI Ractive N 3 stretchesoft he end-on (C 1v )a nd side-on (C 2v )p otassiuma zide isotopologues. [a] C 1v  141414  151414  141415151415141514  1515141415151 51515 the ratio end-on/side-on was about 3/2 whereas after irradiation with UV light (l = 273 nm) it changed to 2/3. While this interconversion waso nly observed in neon, in argon, the side-on KN 3 is the main product after l = 470 nm photolysis. In principle, three reactionm echanisms can be considered for the formationo fm olecular KN 3 in the experiments described above:i )recombination of K + and N 3 À ions, ii)reaction of potassium atoms and N 3 radicals, and iii)reaction of potassium nitrene (KN) and N 2 molecules [Eq. (1)].I nt he neon and argon experiments, however,t he former two reactions can be ruled out since neither N 3 À anions nor N 3 radicals were observed in these matrices. KN is calculated to have at riplet ground state (MRCI(+ Q)), [2] while the reaction of KN + N 2 (1) is assumedt op roceed with excited singlet KN molecules after photoexcitation and intersystem crossing( ISC) to their lowest singlet state. Note that we have not observed the vibrational band of KN (324.4 cm À1 ) [2] in our experiments, because this band is expected to be rather weak and beyond the range of our MCT detector. Nevertheless, KN should be presenta sak ey intermediate for the formation of KN 3 .
In solid nitrogen, large amounts of end-on KN 3 and only small amountso fs ide-on KN 3 were observed after annealing to 25 K. At the same time high amounts of free N 3 À were consumed during this process, suggestingt hat KN 3 might also be formed by ar ecombination of K + and N 3 À ions in solidn itrogen. This observation explains the high abundance of end-on compared to side-on KN 3 in solid nitrogen:The partial negative chargeso nt he two terminal nitrogen atoms in N 3 À favor an electrostatic recombination of K + and N 3 À and the formation of the thermodynamically less stable end-on KN 3 . Figure 5s hows the 14/15 Ni sotopicI Rb and patterns obtained after co-depositiono fl aser ablated KC 14 Nw ith 15 N 2 doped neon (a) and of KC 15 Nwith 14 N 2 doped neon (b) after irradiation with blue light (l = 470 nm). Obviously,s pectra Figure 5a and ba re complementary.D ue to the purity of 98 atom % 15 No f the KC 15 Nused for the experiment,traces of the K 14 N 14 N 14 Nisotopologue can be observed in Figure 5b.T he high yield of K 15 N 3 in Figure 5b indicates that 15 N 2 is formedi ns ubstantial amountsb yr ecombination of 15 Na toms from laser ablated KC 15 N. The absence of any ( 15 N 2 14 N) isotopologues in Figure 5b and of ( 14 N 2 15 N), and K 15 N 15 N 15 Ni sotopologues in Figure 5a implies the absence of ar eaction between K 15 Na nd 14 N 2 producing K 14 Na nd 14 N 15 No rv ice verca. On the other hand, K 14 Nc an react with 15 N 2 not only to form the addition product K 14 N 15 N 15 Nb ut also to form K 15 N 14 N 15 Na nd K 15 N 15 N 14 N. This observation may indicateaK(h 3 -N 3 )i on-pair-intermediate or transition state involvingac yclic h 3 -N 3 À ring in the course of the photoinduced reactiont hat finally rearranges to one of the three possible mixed isotopologues according to Equation (2). Such at ransient cyclic N 3 À hasa lready been postulated by Michl and co-workers. [6b] In addition to the free C 3 À and N 3 À anionsa nd the ion pairs KC 3 and KN 3 the IR spectra of the deposits obtainedf rom IR laser ablation of potassium cyanide also showed bands which can be attributed to well-known free radicals such as the C 3 , [12] N 3 , [6a] CN, [10] NCN, [13] and CNN [14] radicals (for further detailsa nd ad iscussion of the reaction mechanisms see the Supporting Information). 14 N 15 N 15 No rK 15 N 15 N 14 No rK 15 N 14 N 15 In the present paper we have shown that IR laser ablation of potassium cyanidel eads to ac omplex mixture of polynitrogen and polycarbon species. In this study,m olecular KN 3 and KC 3 are reported for the first time. To the best of our knowledge, previous IR spectroscopic studies on potassium azide were performed only on crystalline KN 3 , [15] while severalstudies were reportedo nm olecular group 2, [16] group 13, [7,8] and transitionmetal azides. [17] The ion pair KN 3 exists in an end-on and a side-on form, which are separated by ab arriero fa bout 12 kJ mol À1 at coupled-cluster level. Molecular potassium nitrene (KN) is assumed to be ak ey intermediate produced by IR-laser ablationo fp otassium cyanide(KCN), that reacts photochemically with dinitrogen to molecular potassium azide (KN 3 ). Laser ablationo fK CN could therefore be as uitable route for the gas-phase generation and spectroscopic detection of the elusive KN molecule, which is of interest as ap otentiali nterstellar molecule. Awareness of the photochemical reaction of KN and N 2 might also enable abetter understandingofm echanisms involved in processes of photochemicaln itrogen fixation. Figure 5. IR spectra recorded after co-deposition of laser ablated KCN with N 2 in solid Ne at 6K.D ifference spectra after photolysis (l = 470 nm, 10 min) of a) K 12 C 14 Nw ith 15 N 2 (0.2 %), b) K 12 C 15 Nw ith traceso f 14 N 2 .Bands pointing downwardsi ndicateformation of the corresponding species.