摘要:

Non-porphyrin Photosensitizers in Biomedicine Mark Wainwright Department of Chemistry University of Central Lancashire Preston UK PR 7 2HE 1 Introduction The modern development of photosensitizers in medicine began with the use of porphyrins and ultraviolet light to image neoplastic tissue during surgery. This linked with the fact that some por- phyrins were known to produce phototoxic species made a good argument for the use of porphyrin derivatives in the targeted destruction of tumours and opened the way to the current clinical field of photodynamic therapy (PDT). An excellent review on the use and modern development of porphyrins and porphyrin-based photosensitizers appeared in this journal recently.’ The current article concentrates on what one might consider to be the poor rela- tions of the photosensitizer world viz. the natural and synthetic non- porphyrin derived photosensitizers. Photodynamic therapy was developed as a novel treatment for cancer. The idea of dual selectivity i.e. a tumour-specific drug (the photosensitizer) used in conjunction with fibre optic targeted light delivery is intended to give a treatment modality free from the noxious side-effects encountered with conventional chemotherapy or radiotherapy. In addition the variety of toxic species generated (directly or indirectly) by PDT at the tumour site may be of use in multi-drug resistant (MDR) tumour types -literally swamping the increased defence mechanisms present or attacking the cells via novel routes. The phenomenon of MDR in cancer is quite often encountered and makes the disease even more refractory. Haematoporphyrin derivative as its name suggests comes from a natural source although its use in nature is not as a photosensitiz- ing compound. There are many more examples of natural non-por- phyrin photosensitizers which have evolved over millions of years -normally in plants for chemical defence against microbial or her- bivorous attack. The isolation and elaboration of compounds such as the perylenequinone pigments (PQPs) and the furanocoumarins represent a major advance in this area of research. The requirement for new photosensitizers arose from the fact that the first generation porphyrins used in PDT have several drawbacks. They exhibit poor light absorption properties in the near infra-red the region of the spectrum giving the maximum penetration of light through tissue and thus being of most use in the treatment of tumours. Early porphyrin drugs were not single compounds but mixtures thus making it difficult to establish structure-activity rela- tionships and so improve the efficacy of treatment. In addition from Mark Wainwright was born in Workington Cumbria in 1962. After taking his Chemistry degree at Leicester University he moved to the University of Leeds where he undertook research the time of early clinical trials long-lived skin photosensitivity (typically < 3 months) was apparent. While this side-effect in no way compares with the immunosuppression and general malaise experienced by patients under treatment with conventional anti- cancer drugs it is nevertheless a side-effect which given the current efforts in drug design should be removed in the near future. However second generation porphyrins such as Temoporfin are also known to cause post-treatment skin photosensitivity.’ 2 New Photosensitizers Synthetic photosensitizers certainly prefigured the use of por- phyrins etc. in malignant disease. Raab’s experiments with eosin and acridine for the photodestruction of paramecium at the turn of the century were followed closely by the first clinical application of PDT -the use of eosin against skin cancer -in 1903.* It is interest- ing also to recollect the early use of methylene blue (MB) and light in the treatment of inoperable cancers three years later.3 The idea of porphyrin localisation and its use in tumour staining is also pre- dated by the use of synthetic dyes such as the phenothiaziniums and benzo[a]phenoxaziniums both of which groups (e.g. MB and Nile blue) have been investigated as agents for PDT but thus far have found little clinical use other than in their continuing application in pre-surgical vital staining and histology. During the late 1940s a series of investigations was carried out on the efficacy of synthetic dyes on tumour destruction in animal^.^ The simple rationale behind the research lay in the field of vital staining and it is this work which has inspired much of the current activity on cationic photosensitizers. The use of photosensitizers against other types of disease is a bur- geoning field which again has its origins in the early days of chemotherapy. The use of e.g. toluidine blue in the eradication of the bacteria responsible for dental caries or of the causative organ- isms in oral candidosis can be traced back to the clinical demon- stration of such pathogens using vital stains. If a known synthetic photosensitizer such as toluidine blue can be seen to stain a patho- genic organism sufficiently then it should be possible to destroy that organism via subsequent irradiation. 3 Physicochemical Properties The process of the design and synthesis of new photosensitizers can start from simple ideas. For instance in the preceding example is it possible to synthesise more specific analogues? Can they be made more photoactive? More often than not the answer to both of these questions is ‘Yes!’ However a note of caution must be added. If the photoactivity of a compound is increased by for example halogen substitution it should be remembered that other properties of the molecule will be altered such as its degree of lipophilicity (nor- mally such substitution will increase the log P of the compound see following). This in turn may change the cellular uptake and local- isation of the photosensitizer and possibly its inherent or dark tox- icity. As in any other idea of drug design the various uptake and toxicity parameters must be optimised to give a therapeutically useful compound. It is still a constant surprise to the author that more analogues of commercially available dyes are not produced and tested. Chemicals required for use as photodynamic agents will invari-ably end up in solution either in the dosage form or as a conse- quence of physical metabolism. It is necessary for the medicinal chemist to be able to say how the dissolved photosensitizer will behave in the various pharmacological compartments in which it finds itself. This often comes down to the partitioning behaviour of into photoconductive polymers. After his PhD (/988) he remained in Leeds joining the nascent Centre for Photodynamic Therapy under Professor Stan Brown synthe- sising new photosensitizers for PDT. He is currently a Senior Lecturer in Organic Chemistry at the University of Central Lancashire and a Visiting Specialist in Medicinal Chemistry ut the Royal Preston Hospital with research inter- ests in photodynamic anti-cancers and antibacterials. 351 Lipophilic species Log P +ve octan-1-01 WaterHydrophilic species Log P -ve Figure 1 The partitioning behaviour of a photosensitizer the compound between the water which is quite ubiquitous in the biological milieu and the various types of lipids encountered for example at the cell membrane or as part of the complex mixture of blood proteins i.e. the hydrophilicAipophilic balance of the drug. A simple in vitro measure of this behaviour is taken as the logarithm of the partition coefficient of the compound between a two-phase mixture normally water and octan- 1-01(see Figure 1). log P = log {COct/Cw} A problem particularly relevant to new photosensitizers based on commercial dyes arises here in that they are invariably used as charged species. Thus dyes such as MB and toluidine blue are nor- mally encountered as salts with the dye chromophore being the cation. In solution both dye cations are hydrophilic (log P = -0.9). However in biological systems toluidine blue is partially converted into a neutral form by deprotonation and both dyes are liable to metabolic reduction which also renders them as neutral species. In each case the neutral form is highly lipophilic (log P > +3). This behaviour is not apparent from simple water-octanol partitioning and so great care must be taken to avoid too much extrapolation from the results of such in vitro work. The light absorption properties of putative new photosensitizers are also obviously of great importance (this is in fact the case with any drug overlooking the photosensitizing activity of new agents can lead to needless side-effects such as skin sensitization). Visible and ultraviolet light are absorbed by various components of biolog- ical systems and this endogenous absorption may be critical in the use of photosensitizers if the wavelengths used for irradiation are too short. The natural pigment melanin has a wide absorption which covers most of the visible region thus facilitating its protective capability against solar radiation. Haemoglobin has several absorp- tion bands in the visible region. In addition tissue scattering of light is important at shorter wavelengths. Thus in tumour work new photosensitizers are normally designed to absorb at long wave- lengths -the normal range given is 600-1000nm. This is known as the ‘therapeutic window’. Photosensitizers which are useful against bacteria viruses etc. i.e. in external eradication often absorb at wavelengths below 600 nm. The efficacy of the photodynamic action associated with new photosensitizers should obviously be high if any advance is to be made on currently available agents. In terms of in vitro testing certain performance indicators may be used,among which the singlet oxygen quantum efficiency (@J is often important since it is intimately involved with the effectiveness of Type I1 processes (see Figure 2). It has often been demonstrated that the replacement of an atom in a lead structure with one of higher atomic number leads to an increase in the @A value. This is known as the internal heavy atom effect. In electronically excited molecules spin-orbit coupling facili- tates intersystem crossing (ISC Figure 2) allowing otherwise for- bidden changes in the spin state (i.e. singlet-triplet). Since the spin-orbit coupling constant is proportional to the fourth power of the atomic number of the element concerned the presence of a ‘heavy atom’ (e.g. bromine or iodine) in a molecule enhances the degree of spin-orbit coupling. In terms of the compounds discussed here this should lead to an increase in the triplet yield and improved photosensitizing activity. CHEMICAL SOCIETY REVIEWS 1996 Type I mecbaaism Redox reaction with biomolccules production of superoxide.etc. Type I1 reactions Mediated by singlet oxygen e.g. lipid peroxidation.guanosine hydroxylation hm I kj rl / Type I 3Ps + Sub -Ps+. + Sub-. (electron transfer) Sub-. + 30,-Sub + 0;. -HO-+ HO. Type I1 3Ps + 30,-IPS+ + ‘0,(energy transfer) lo + Sub -Sub-OOH (peroxides etc.) (Where Ps = photosensitizer; Sub = substrate or solvent) Figure 2 Modified Jablonski diagram showing the various photoprocesses involved on excitation of a dye molecule. Key So-singlet state SI -first singlet excited state; ISC -intersystem crossing; TI -first triplet excited state; F -fluorescence P -phosphorescence IC -internal conversion. It is however dangerous to take improved values as a guar- antee of high photodynamic activity -there are many other factors involved in moving from in vitro chemical testing to cell culture and similarly to mammalian systems. Also the reverse may be true an apparent lack of photosensitizing ability does not necessarily mean that the compound undergoing testing will be ineffective in cells (e.g.Victoria blue BO). A brief outline of Type I and I1 pho- toprocesses is given in Figure 2. 4 Naturally Occurring Photosensitizers and Their Derivatives 4.1 Psoralens Psoralen derivatives have been used for thousands of years in the East and Middle East in the treatment of various skin disorders but more recently the structure-activity relationships and sites of action of the psoralens have been elucidated and synthetic analogues pre- pared. The field of psoralen photomedicine and in particular pho- tophoresis has become incredibly active in recent years with a bewildering array of newly synthesised compounds as well as the in vitro testing of natural congeners and structural isomers of the furanocoumarin unit. A representative sample of the more promis- ing compounds appears in Figure 3. The concept of photophoresis is based on the selectivity of pso- ralen derivatives for malignant cells such as the lymphocytes impli- cated in cutaneous T-cell lymphoma. Oral administration of the psoralen drug leads to its uptake by malignant T-cells in the blood- stream. Sequential removal of 0.5 1 aliquots of blood component separation and illumination of the white cell fraction with the appro- priate wavelength of ultraviolet light leads to DNA damagevia pho-toadduct formation and thus to direct cytotoxicity. Protein damage at the cell membrane leads to cell death or causes sufficient changes in cellular morphology such that the reintroduction of the white cell fraction into the bloodstream causes an autovaccination effect i.e. the malignant cells are not recognised by the body’s immune system and are thus de~troyed.~ There are many examples in early photochemotherapy which suggest that simple psoralens cause cross-linking in DNA. This might be deleterious in the long term and merits special attention if the disorder under treatment is non-malignant. Psoralen functional- isation in the pyrone ring can yield compounds which are unable to form cycloadducts via this ring due to steric factors. Since methy- lation in the furanocoumarin nucleus furnishes compounds e.g . NON-PORPHY RIN PHOTOSENSITIZERS IN BIOMEDICINE- M WAINWRIGHT A‘ 54 54 5 1‘ m 1 z 1‘0p-1 0 54 Psoralens Angehcins W X Y Z 0 CH 0 0 Psoralen NH CH 0 0 Pyrrolocoumarin S CH 0 0 Thienocoumarin 0 N 0 0 8 Azacoumarin 0 CH NH 0 Furano-2 quinolone 0 CH 0 S 2-Thiofuranocoumarin 0 CH 0 Se 2 Selenofuranocoumarin Figure 3 Psoralens angelicins and bioisosteres 4,8,5’-trimethylpsoralen or 4,6,4’-trimethylangelicin which cause DNA cross-IinkingP the steric factor is obviously important The 3- ethoxycarbonyl analogue is sufficiently hindered to give rise only to mono-adducts with DNA Psoralens can also cause hydroxylation of guanosine in nucleic acids a mechanism often associated with the intermediacy of singlet oxygen However in recent work with 3-ethoxycarbonyl-psoralen which is efficient in causing guanosine hydroxylation neither degassing of the reaction media nor the use of deuteriated water gave results consistent with a Type I1 mechanism It is possi- ble though that the effective intercalation of the furanocoumarin chromophore with DNA may make it difficult for D,O interaction to occur In many ways the idea of psoralen -protein interactions has been overlooked in recent years yet the widely reported psoralen -DNA adduct formation cannot explain the immunotherapeutic basis of the treatment Although it is suggested that psoralen photodamage to the cell membrane may be due to psoralen photoadducts with biomolecules other than nucleic acids this sort of damage can certainly be envis- aged as occurring via a photodynamic rather than a photo-chemotherapeutic route since the production of singlet oxygen by psoralen derivatives in solution is now well established If this is indeed the case the ongoing efforts in new drug design and synthe- sis in this area and particularly those involving increasing the @A values are well justified When a drug is established for the treatment of a particular dis- order analogues are synthesised as a matter of course in order to improve on the activity of the lead compound This is in many cases the raison d’gtre of the medicinal chemist There are two main reasons for the development of analogues of psoralen Hypencm Perylenquinonoids R R1 WR‘ CH,C( OH)MeCH( COMe) WR’ -CH,CMe=C(COMe) -CH,C(OH)Me -CH,C(OH)Me CH,CHMeCOPh CH,CHMeOCO-p C,H,OH Figure 4 Hypericin and the perylenequinonoid pigments increasing the selectivity for malignant cells or increasing the pho- toactivity at the target site for example via the heavy atom effect Thus many bioisosteres and structural isomers of the fura- nocoumarin nucleus have been isolated or synthesized de novo (see Figure 3) The replacement of the furan oxygen with sulfur or sele- nium gives rise to compounds having much improved photoactiv- ity In addition the 8-azapsoralens -i e analogues arising from replacement of carbon-8 with nitrogen -exhibit lower incidences of DNA cross-link formation The activity of 4,4’ ,S’-trimethyl-8-azapsoralen in terms of its inhibition of DNA synthesis in Ehrlich cells has been reported to be six times that of 8-methoxypsoralen (8-MOP) and it was efficacious in the clearance of psoriasis in recent clinical trials However 8-MOP showed slightly higher activity than the azapsoralen The evidence of a decreased inci- dence of cross-linking by the azapsoralen compared to that by 8- MOP should encourage further clinical testing 42 Anthracyclines Insofar as the use of conventional anticancer drugs in PDT is con- cerned it may be possible to exploit a favourable combination of tumour selectivity and reasonable light absorption properties to improve on typical antitumour activity Likely candidates fulfilling such criteria are members of the anthracycline group such as Doxorubicin Synthetic anthraquinone textile dyes are well known as causative agents in skin phototoxicity e g ‘bikini dermatitis’ Drugs such as Doxorubicin (Figure 4) have also shown phototoxicity in vitro although to a lesser degree than mainstream PDT agents Additionally the A, values for such drugs are usually not much greater than SO0 nm These factors may not be too important since the current view on anthracycline PDT is to use the photoactivity as an additive therapy and not as a single means of tumour destruc- tion Although drugs like Doxorubicin show a reasonable selectivity for turnours they also exert adverse effects on healthy cells in par-ticular cardiotoxicity The utilisation of the extra photodynamic activity associated with such drugs may make it possible to produce the same levels of antiturnour activity at lower drug doses with a concomitant decrease in side effects The likely clinical inception of this type of combinative therapy must be given extra momentum by the fact that through long clinical use the pharmacology of the anthracycline drugs is by now very well understood 43 Hypericin and Hypocrellins Herbivorous animals for example domestic cattle have long been known to suffer from a skin disorder (hypericism) after the inges- tion of a photosensitizer contained in the weed St John’s Wort (Hypericurnperforaturn) Skin photosensitization occurs due to the HOtJNY Doxorubicin R2 R3 OMe OMe Hypocrellin A OMe OMe Hypocrellin B R2/R3 CH,OCH Cercosporin OMe OMe Calphostin C transport of the pigment hypericin in the animal's bloodstream to the epidermal capillaries and subsequent activation by sunlight Similar findings in China led to the use also of extracts of Hypocreffa barnbusae in the traditional treatment of skin disease The perylenequinonoid pigments (PQPs) and the related hypericin implicated here were investigated and their favourable photoprop- erties -near infrared absorption and high singlet oxygen efficiency (typically@A ca 0 8) -have promoted their in virro testing for PDT Hypericin and the PQPs have been shown to be pharmacologi- cally active in terms of malignant disease Their high log P values coupled with formal negative charge explain the lipoprotein binding and membrane targeting associated with the pigments In addition the large pseudoplanar area makes them obvious candidates as antivirals -indeed hypericin is undergoing clinical trials in AIDS patients Increased interest in hypocrellins has arisen due to the fact that they have been found to be inhibitors of protein kinase C (PKC) a key enzyme in the proliferation of tumour cells Semi-synthetic approaches to new hypocrellin photosensitizers involving their action against PKC were made starting from the known PQP cer- cosporin via conjugate addition of e g ,thiophenol at positions 5 and 8 Calphostin C (see Figure 4) appears to be more active in PKC inhibition than any of the newly tested derivatives and it is sug- gested that the increased photoactivity against PKC arises via the addition of cysteine residues in active sites of the protein (through -SH) at positions 5 and 8 of calphostin C Thus when these posi- tions are blocked the photoactivity decreases lo Several new derivatives of the hypocrellins have exhibited higher photoactivity in cell culture The hypocrellin structure is such that the phenolic groups (positions 4 and 9) may be replaced by nucle- ophiles and alkylamino substitution has recently been shown to have dramatic effects on the phototoxic nature of the prepared derivatives Thus the 4,9-bis(butylamino)-derivative of hypocrel- Iin B has a much higher E, value than the parent compound leading to far greater phototoxicity against EMT-6 cells in culture At the same time the introduction of the butylamino side chains decreases the dark toxicity of the derivative relative to the parent thus yielding a compound reportedly having a high lightldark toxi- city ratio (2167) This study has demonstrated the potential of the hypocrellin skeleton for functionalisation leading to changes in physicochemcial properties and improved characteristics for PDT II 5 Synthetic Photosensitizers 5.1 Cyanines Cyanine dyes have been used as photosensitizers from the early days of the photographic industry In addition work carried out by Carl Browning at the end of the First World War clarified some structure-activity parameters for a large series of the dyes as antibacterials The structure of cyanine dyes is such that analogue formation is straightforward e g via variation in heteroatom N-alkylation length of polymethine chain etc ,and has led to a large number of compounds being synthesised Research activity in this area has therefore been considerable Generally the A, for a given cyanine dye may be increased by lengthening the polymethine chain This may however lead to problems with photoisomerisation a major deactivation route for cyanine dyes Increasing the chain length often yields dyes which are less chemically stable and causes decreased aqueous solubility A careful balance of these factors is thus required when designing likely candidates for testing as photosensitizers in biological systems This work has resulted in the emergence of such dyes as Merocyanine 540 (MC540) EDKC (a kryptocyanine) and thence chalcogenapyryliums as active agents in the photodynamic treat- ment of malignant disease Whereas EDKC and its congeners appear to have progressed little in recent years MC540 is particularly useful against leukaemic cells and has undergone phase I clinical trials for the purging of autologous bone marrow grafts MC540 was in the vanguard of the synthetic non-porphyrin photosensitizers evaluated CHEMICAL SOCIETY REVIEWS 1996 in the mid-late 1980s,and as such the understanding of its mode of action -and therefore that of other cyanine photosensitizers -is now extensive MC540 may be used in two different ways against target cells The conventional mode of action is that of membrane photooxida- tion the lack of intracellular targeting reflecting the anionic char- acter of the photosensitizer Conversely the idea of preactivation of MC540 uses photoirradiation of the photosensitizer prior to its delivery to the tumour erc the reaction of the dye with singlet oxygen yielding stable photoproducts such as merodantoin and merocil (Scheme 1) These photoproducts are known to act against topoisomerase 11 an enzyme intimately involved in DNA replica- tion and thus to inhibit macromolecular synthesis by conventional chemotherapeutic routes One advantage of this technique lies in the lack of toxicity of the photoproducts to healthy cells allowing attack on e g metastatic disease since no further light delivery is required MC540 i 0 Meroxazole Merocil Merodantoin Scheme 1 Preactivation photoproducts of MC540I2 The structure of MC 540 allows wide scope for the design of analogous series This is necessary due to the low @A value (0 002) of the dye On utilising the heavy atom effect in substituting sele- nium for sulfur at position 2 of the barbiturate moiety a significant increase in the quantum yield of singlet oxygen is produced Photoisomerisation of the central double bond of the polymethine chain a major deactivation pathway for MC540 is also absent in the seleno-analoguel3 Both of these factors obviously contribute to the increased (DA value in vitro (see Figure 5) However photobleach- ing of the analogues in the cellular milieu under experimental con- ditions means that MC540 is still likely to be a better antileukaemic Carbocyanines and related dyes have also been proposed as PDT agents As with MC540 their levels of activity suffer from photoisomerisation This problem has been approached in two ways the introduction of substituents into the rneso-position of the NaO3S' I I CsHB GHB X R Amax'nm @'-l(llpo5'>mes) Q3,EtOH)Merocyanines 0 (MC540) 535 0 001 S 555 0 002 Se 56 1 0 38 Dicarbocyanines S Et 551 0 01 0 001 Se H 574 0 08 0 014 0 (DHOCI) H Figure 5 Cyanine photosensitlzersl3 l4 NON PORPHYRIN PHOTOSENSITIZERS IN BIOMEDICINE-M polymethine chain or the use of long chain alkyl groups on the 3,3’-positions While the former approach rigidifies the polymethine chain directly it has been demonstrated in liposomes that the long chain N-alkyl substituents cause physical anchoring and thus rigidi- fication of the molecule within the lipid bilayer I4 In addition (D values may be increased via the heavy atom effect (see Figure 5) Compared to MC540 the different charge on the carbocyanines obviously has ramifications regarding sites of action Thus whereas MC540 acts at the plasma membrane cationic cyanines may be internalised typically in the mitochondria However as more ex- amples are synthesised new targets are being identified For example 3,3‘-dihexyloxacarbocyanine iodide (DHOCI -see Figure 5) causes specific photodamage to microtubules Is Such results are very promising given the great cellular sensitivity of the target site 5.2 Phenothiazinium Photosensitizers The phenothiazinium dye methylene blue (MB) has been used extensively for over a century as a vital stain It is widely employed in the clinical diagnosis of a variety of diseases and as a tumour marker in surgery Its use as a PDT agent IS not widespread which might be surprising given the extensive commercial use of MB as a photosensitizer were it not for the facile reduction of the dye in the biological milieu MB is reduced to leuco methylene blue (LMB) by the ubiquitous cellular coenzymes NADH and FADH In point of fact the MB + LMB conversion forms the basis of the tuberculin test in milk The reduction of MB by tumour cells means that the efficacy of photosensitization is decreased LMB is colourless and is thus inac- tivable by the long wavelength light used in PDT In addition the pK value of LMB is low (5 8) compared with MB leading to a low level of ionisation of the reduced species (3 1% ionised at pH 7 3) High ionisation is essential for efficient DNA intercalation and photodamage to DNA is thought to be important in the photocyto- toxicity of MB There are several closely-related commercial analogues of MB (Figure 6) the demethylated azure stains and thionin and nuclear- + R R‘ R2 R’ R4 R“ Me,N Me,N H H H H H H H H H H Me,N H H H H H MeNH H H H H H H2N H H H H H Me,N EtNH H Me H H H H H Me H NO Me2N H Me Me H H Me,N H H H Me H WAINWRIGHT substituted derivatives such as toluidine blue 0 (TBO) methylene green and Taylor’s blue The majority of biological and clinical work has involved MB and TBO once again reflecting their wide- spread use in vital staining Although as mentioned above MB is not yet widely used in clin-ical PDT its efficacy in the treatment of several tumour types has been demonstrated both in animals and recently in patients Recurrent inoperable oesophageal tumours have been treated with locally delivered MB and were illuminated with a modified optical fibre intraluminally giving tumour eradication l6 In these cases the photosensitizer was injected directly into the tumour thus further minimising side-effects One of the reasons for the apparent reluctance in the use of MB-PDT could be the dark toxicity associated with the photosen- sitizer To a certain extent this may be overcome by physical selec- tivity -1 e applying the photosensitizer directly to the tumour site as in the preceding case This also explains the trial use of MB in bladder cancer since the photosensitizer solution can be instilled into the bladder cavity quite simply Extra selectivity in anticancer drugs can be achieved using anti- body labelling This has been attempted with MB by the synthesis of side-chain maleimido- and succinimido- derivatives The result- ing MB-protein conjugates are reported to be slightly less photoac tive in vrtro than the parent compounds The above example aside there are few reports of novel synthetic (1 e non-commercial) analogues of MB This may be due to the mis conception that such dyes may only be produced using archaic recipes from the late 19thcentury dye industry While it is true that the more complex (ring-substituted) derivatives are best synthe sised via dichromate oxidation of diaminoaryl thiosulfonic acids the simpler 3,7-bisamino-analogues of MB are easily obtained through 5-phenothiazinium iodide’* (Schemes 2 and 3) and using this method it is possible to achieve differing amino character However the utility of the ‘old’ method is shown by a recent example prepared quite simply from 1 ethyl-2,2,4-trimethyl-1,2,3,4-tetrahydroquinoIineThe reduced quinoline derivative can be thought of as an elaboration of the N,N-dimethylaniline used routinely in the industrial preparation of MB itself The resulting + MBD Amax’nm Methylene Blue 661 Azure A 633 Azure B 648 Azure C 616 Thionin 598 Methylene Green 656 New Methylene Green 630 Taylor’s Blue 649 Toluidine Blue 626 X A,,,,’nm @A O‘ 632 0 005 S 652 0 025 Se 659 0 650 Replacement of NHEt with NH = Nile Blue A Replacement of NHEt with H and NEt with NME Meldola s Blue Figure 6 Phenothiazinium and Nile Blue der~vatives~~ %N + “2 Phenothlazmiums Benzo[a]phenothiazinlums Scheme 2 Synthesis of derivatives Methylene Blue and Nile Blue H Scheme 3 Synthesis of Methylene Blue derivatives from phenoth~azine’~ pentacyclic MB derivative (MBD -see Figure 6) showed selective uptake by a fibrosarcoma in rats but was not taken up by metastatic neoplastic cells I9 The use of MB against viruses is a rapidly expanding area The photoactivity of the dye against highly infective enveloped viruses such as HIV and hepatitis C is important because of its application in the disinfection of donated blood The disinfection technique is quite straightforward the long wavelength absorption of MB means that blood mixed with a small amount of photosensitizer may be illuminated while still in the ‘blood bag’ Although MB can inacti-vate free viruses without collateral damage to other blood compo-nents such as proteins platelets etc ,there is a problem in that intracellular viruses remain unaffected The answer may lie in a PDT/photochemotherapyapproach-i e using MB to eradicatefree viruses in the sample in conjunction with a psoralen-type agent for intracellular viruses 2o This method would obviously require more sophisticatedlight delivery 53 ToluidineBlue (TBO) Similar in structure to MB TBO has similar properties as regards DNA intercalation although the A, of TBO is rather shorter at 626 nm (in water) which obviously makes it less attractive in terms of light penetration in tissue However,TBO has found widespread use in the diagnosis of oral disease since it is a selective stain for oral cancer and also for various oral pathogens This has led to an investigation into photosensitizer-mediated photodestruction of bacteria and other microbes pertinent to modern dentistry by Michael Wilson and coworkers at the Eastman Dental Institute in London CHEMICAL SOCIETY REVIEWS 1996 The presence in dental plaques of bacteria leads to the erosion and disease of the oral surface involved -either tooth enamel (dental caries) or gum tissue (gingivitis) Chronic disease requires the removal of a large amount of enamel and dentine or tissue in order to eradicate the bacteria The use of a pathogen-specificpho-tosensitizerand low-powerlaser light to destroy the bacteria would thus entail less removal of healthy tissue This phototoxicity has been demonstrated by TBO for a wide range of oral pathogens such as Streptococcus mutuns and Lactobacillus casei i The technique is also of use against oral Cundida stains -important pathogens in terms of the treatment of AIDS patients -and has been demon-strated against the clinically important methicillin-resistant Staphylococcus aureus (MRSA) 22 MRSA infection,particularly in post-operative situations (biofilm infestation of implants catheters etc ) is an increasing cause of morbidity Another refractory stain of pathogens,Helicobacter which has in recent years received a good deal of attentiondue to its reputed involvement in gastric ulcers and cancer is also susceptibleto photodynamic treatment using TBO 23 Although the photokilling of bacteria and yeasts has been widely reported the clinical importance of this work is that it has been demonstrated in biofilms since this is the most likely pathogenic presentation The reason that drug resistance is overcome using the photodynamic approach must surely be the same as in resistant tumour cells -a different type of toxicity means that the mech-anisms available to combat it are absent Other types of photosen-sitizer have been shown to be effective in this respect such as aluminium disulfonated phthalocyanine,however this is less effec-tive than TBO 21 The high activity of TBO and MB against yeasts relative to established tumour PDT agents such as aluminium disul-fonated phthalocyanine and dihaematoporphyrin ester was evident in the in vitro photodestruction of Candida albicans24 This is perhaps unsurprising in the latter case since tumour phototoxicity due to dihaematoporphyrin ester is normally associated with its localisation in tumour vasculature,not intracellularly 5.4 Nile Blue Analogues Nile blue is taken up extremely well by tumour cells thus making it an excellent tumour marker It is actually a benzo[uJphenox-azinium salt synthesised from 5-(N,N-diethylamino)-2-nitrosophe-no1 and 1-naphthylamine Simple analogues with different amine functionality in the 5-and 9-positions may be obtained easily from Meldola’s blue (Figure 6) or derivatives thereof Nile blue and several of its analogues were examined as potential photosensitiz-ers for PDT but very low singlet oxygen efficiencies preclude their usage The reason for the poor efficiency is known to be the low occurrence of intersystem crossing (see Figure 2) This has been remedied by introducing heavy atoms into the chromogen Indeed the recent work on benzol ujphenoxazinium chalcogen analogues stands as an excellentexample of the heavy atom effect (see Figure 6) Nile blue analogues containing sulfur at position 7 are synthe-sised via a similar method to that for MB except that the 4-dialkyl-aminoaniline-2-thiosulfonicacid is isolated purified and then oxidised (dichromate) together with a 1-aminonaphthalene (Scheme 2) The nile blue derivatives so produced were found to have greatly increased singlet oxygen efficiencies and also to be lysosomotropic -i e to accumulate preferentially in lysosomes The presence of the 9-amino (in later derivatives 9-ethylamino) group allows the formation of the neutral quinoneimine species at physiological pH (see Scheme 4),thus facilitating crossing of the lysosomal membrane Once inside the organelle lower pH causes protonation of the quinoneimine blocking the reverse passage Photodestructionof the lysosomes releases hydrolytic enzymes into the interior of the cell thus mediating cell death The activity of 9-ethylamino derivatives was further enhanced by inclusion of sele-nium at position 7 again increasing the singlet oxygen efficiency In the cultured cells used (EMT-6 mouse mammary sarcoma) the seleno-analoguewas I000 times more phototoxic than photofrin I1 In addition It has been shown that reduction and indeed photore-duction of the photosensitizers occurs in the absence of oxygen yielding the neutral leuco species (Scheme 4) 25 NON-PORPHYRIN PHOTOSENSITIZERS IN BIOMEDICINE-M WAINWRIGHT Reduced form 4 Quinoneimne (EtNBS R = R = Et) Scheme 4 Pharmacologically relevant species from Nile Blue derivativesz5 In animal tumour models 5-ethylamino-9-diethylamino-benzola lphenothiazinium chloride (EtNBS) has been shown to be an efficient PDTdrug Thereasonsforthislie,mostprobably,inthe redox activity of the compound in the tumour and in pharmacologically dif- ferent compartments such as the skin Thus subcutaneous adminis- tration of the drug followed by photoirradiation gave tumour destruction without subsequent skin photosensitization because of differences in distribution (eight hours post-injection EtNBS was present in the ratios 4 1 tumour surrounding muscle and 8 1 tumour skin) and also because of the differing metabolic rates of the organs involved The tumour contained mainly oxidised EtNBS whereas the drug is evidently reduced to the inactive leuco-EtNBS in the skin In addition the lysosomal selectivity of EtNBS ensures that it is present inside the tumour cells thus leading to efficient tumour kill rather than destruction of its vascular support The first reported photodestruction of larger-scale murine tumours (diameter 8-1 0 mm) has been achieved using a combina- tion of EtNBS and a benzoporphyrin derivative (BPD-MA) It is thought that the PDT effect of the two drugs is synergistic rather than additive since increasing the light dose or the concentration of either of the drugs singly had no significant effect (and indeed dou- bling the dose of BPD-MA caused death in three-quarters of the animals tested) Examination of the localisation of the drugs showed that BPD-MA was quite diffusely localised throughout the tumour but mainly at the cell membranes and so did not compete with EtNBS 27 55 Rhodamines The discovery that the commercial dye Rhodamine 123 (Rh123) is taken up specifically by mitochondria together with its known use as a fluorescent indicator led to its use as a fluorescent probe in sub-cellular studies Subsequently and with the modern evolution of PDT this was extended to the investigation of Rhl23-treatment of animal tumours Rh123 has a high fluorescence quantum yield (Of = 0 9) and not surprisingly shows considerable dependence for cell- killing on high levels of tumour cell oxygenation 2s Analogue prepa- ration has therefore been carried out to improve on this situation The 4,5-dibrominated analogue of Rh 123 is simply prepared via direct bromination at room temperature Differing ester functional- ity has also been included in recent work to improve cellular photo- sensitizing ability (see Figure 7) although there is a concomitant increase in dark toxicity with the longer chain (butyl) ester presum- ably due to the higher lipophilicity of this compound leading to a lower rate of efflux from cells 29 Tetrabromorhodamine 123 (TBR I23) having two more bromines attached to the parent nucleus is not surprisingly considerably more lipophilic and is found in hydrophobic regions of malignant cells 3O Interestingly when used against multidrug resistant (MDR) cell lines photoinactivation of the cells was not observed but the cellular detoxification agent P-gly- coprotein was inhibited allowing increased uptake of other toxic compounds such as Rh123 or the conventional anticancer drug R R' R* R3 @A Rh123 H H H Me 001 DBR123 H H Br Me 047 TBR123 H Br Br Me 070 Figure 7 Rhodamines Daunorubicin This points to the use of such compounds as TBR123 in the adjuvant therapy of MDR disease either with standard chemotherapeutic agents or with other photosensitizers The main drawback in the use of rhodamines for the PDT of solid tumours is the relatively short wavelength of absorption available Rh123 absorbs light at just above 500 nm This value may be increased by the usual means e g alkyl substitution at the amino groups (rhodamine 6G A, 528 nm rhodamine B A, 543 nm) or by direct halogenation of the chromogen (e g TBRl23) As men- tioned previously this alteration of the parent compound may decrease uptake or introduce unwanted dark toxicity However their proven in virro activity against malignant cells strongly indi- cates the use of rhodamines in the eradication of leukaemic cells from bone marrow extracts in preparation for transplantation 29 5.6 Triarylmethane Photosensitizers The antimicrobial activity of the cationic triphenylmethane dyes is well established and several examples were among those examined for antitumour activity in the 1940s The structure of the triph- enylmethanes is such that the energy accrued from photoexcitation can be dissipated easily via internal conversion (see Figure 2) thus making it difficult to examine parameters such as singlet oxygen efficiencies However this has not precluded examination of tri- arylmethanes for use In PDT and related areas The Victoria blues are well-known commercially available dyes and are similar in structure to crystal violet one of the phenyls being replaced by a 1 -naphthyl group with a secondary amino function at position 4 (see Figure 8) The Victoria blues were included in the R Victona Blues (VBBO R = Et R = NHEt) Q Crystal Violet Figure 8 Triarylmethanes Lewis study4 and exhibited uptake by mammalian tumours This combined with its long wavelength absorption (612 nm) justified a re-examination of VBBO as a photosensitizer Although no evi dence of singlet oxygen production has been found in in vitro tests electron paramagnetic resonance has been used to demonstrate that the dye can photosensitize the production of superoxide 31 VBBO has been shown to be highly photoactive against several tumour cell lines with a low level of dark toxicity the combination of de- localised positive charge and high lipophilicity ensuring mitochon- drial uptake The dye also exhibited very high photoactivity against two human leukaemic cell lines with 99% cell killing at a photo- sensitizer dose of 0 1 kmol dm-3 72 In terms of structure-activity relationships in the Victoria blue series the naphthyl moiety is important for photoactivity since triphenylmethane dyes are generally inactive in this respect Also compounds containing only tertiary amino functionality are less active both in terms of uptake and photosensitizing ability thus sug- gesting that the secondary amino group is involved in drug action as is the case with the Nile blue derivatives (vide supra) The pres- ence of the secondary amino group in the commercial Victoria blues allows simple conversion of the dye cation to the neutral quinoneimine species the Homolka base and this has been used to explain the high activity of VBBO compared with analogues having all-tertiary amino character 33 5.7 Acridines Acridines are among the most widely investigated heterocyclic compounds in the modem history of chemistry Their use extends once again from Ehrlich’s pioneering work in the late nineteenth century through the use of simple aminoacridines in wound anti sepsis during the latter part of World War One and the widely used antimalarial Mepacrine up to conventional present day anticancers such as Amsacrine Acridines were also among the first synthetic heterocyclic dyestuffs In terms of photosensitizing activity acridine was shown to have a photodynamic effect by Raab nearly a century ago as has been mentioned The aminoacridines have long been used as fluorescent probes their cellular ‘targets’ depending unsurprisingly on struc- ture Thus proflavine is an excellent nucleic acid probe whereas its bis(dimethy1amino)-analogue acridine orange accumulates in lyso-somes Alkylation of the ring nitrogen in acridine orange with an ethyl group or larger gives rise to mitochondria1 localisation Given the long medicinal and dyestuffs pedigree of the acridines -and it is known that the DNA strand-breaking activities of a number of commercially available intercalating acridines are enhanced on ph~toirradiation~~ -it is strange that this combination has not yet given rise to much activity in the field of PDT Until recently only acridine orange had been shown to cause tumour pho- todestruction in animals using an argon laser3s Because of the highly fluorescent nature of such aminoacridines it might be expected that singlet oxygen efficiencies are low and indeed there appears to be a strong oxygen tension dependence However lyso- soma1 destruction has been demonstrated using acridine orange and blue light in cultured tumour cells 36 For the photodynamic therapy of larger tumours in vivo,the short wavelengths of absorption of commercial acridines such as acridine orange proflavine acridine yellow etc might be problematical since they fall far short of the therapeutic window The extension of acridine absorbance into the near infrared may be achieved by cyanine synthesis e g 9-styrylacridines (see Figure 9) By the facile condensation of 9-methylacridine with N,N-dialkylaminobenzalde-hyde derivatives so-called acidochromic dyes are produced These in addition to absorbing light at ca 700 nm can exist in neutral or protonated (N-10) forms at approximately physiological pH since they have pKa values of ca 5 The theory behind the use of such compounds in tumour work uses the pH drop in tumours This would give a higher degree of protonation inside the tumour thus adding to the selectivity of the system since the protonated dye is the photoactive species and has an absorption which is bathochromically shifted by ca 200 nm Extrapolating to the idea of tumour treatment in vivo,this would mean that peritumoural CHEMICAL SOCIETY REVIEWS 1996 X R R’ X h,,,,lnma Proflavine NH H H 456 Acridine Orange NME H H 489 Acridine Yellow NH Me H 442 StyryP H H p-Me,NC,H,CH=CH 618 (L EtOH/HCl 436 nm in neutral EtOH3’ Figure 9 Acridines tissues would contain mainly neutral dye which would of course remain unaffected by near infrared light The only drawback to this (preliminary stages) work is that the phototoxicity of the acridines so far synthesised is likely to be low (<5% of the photosensitizing activity of MB zn vitro)37 6 Conclusions Thus far the majority of photosensitizers used in cancer PDT clin- ical trials have either been porphyrins or their derivatives This IS due mainly to the fact that modern PDT has evolved from the natu- rally derived porphyrins such as HpD and DHE It is not surprising therefore that the second generation photosensitizers are based on porphyrins and phthalocyanines Whilst it has been proposed that established conventional chemotherapeutic agents such as the anthracyclines have sufficient photosensitizing ability to allow their use in clinical PDT the avail- ability and well-established chemistry of commercial non-por- phyrin dyes and photosensitizers has also led to their testing as PDT agents zn vitro Using commercial and natural photosensitizers as lead compounds a great deal of design and development has been carried out by various groups in synthesising improved derivatives Several candidates have shown higher levels of activity than por- phyrin-derived materials in cell culture and in animal testing (e g the benzo[ alphenothiazinium derivatives) In addition a wide range of vital intracellular targets has been demonstrated for example the use of DHOCI against microtubules which should lead to higher therapeutic efficiencies than conventional membrane phototoxicity So far however clinical usage of such new compounds has been at best minimal In terms of alternative targets to cancer e g bacteria and viruses there is less rationale for the use of porphyrin species The field of vital staining which has developed over the past century utilises many commercially available dyes As many of these dyes exhibit photosensitizing activity it is logical to suggest their use in photo-dynamic pathogen eradication It is to be hoped that the demonstra- tion of the photosensitizing capability of the non-porphyrin dyes against microbial and viral targets will lead to increased interest from industry 7 References 1 R Bonnett Chem Soc Rev 1995,19 2 Von Tappeiner Munch Med Wochenschr 1903,47,2042 3 A Jacobi J Am Med Assoc 1906,47,1545 4 M R Lewis P P Goland and H A Sloviter Anal Rec ,1946,96,201 5 1 M Schmitt S Chimenti and F P Gasparro J Photochem Photohiof B 1995,27 101 6 A Guiotto A Chilin P Manzini F Dall’Acqua F Bordin and P Rodighiero Farmaco 1995,50,479 7 W G Warner W C Timmer R R Wei S A Miller and A Kornhauser Photochem Photobiol ,1995,61,336 8 R E Saxton M B Paiva R B Lufkin and D J Castro Sem Surg Oncol 1995,11,283 9 G Lavie Y Mazur D Lavie and D Meruelo Med Res Rev 1995,15 111 10 Z Diwu J Zimmerman T Meyer and J W Lown Biochem Pharmacol ,1994,47,373 NON-PORPHY RIN PHOTOSENSITIZERS IN BIOMEDICINE-M WAINWRIGHT 11 E P Estey,K Brown,Z Diwu,J Liu,J W Lown,G G Miller,R B Moore J Tulip and M S McPhee Cancer Chernother Pharmacol 19%,37,343 12 R Sharma L Arnold and K S Gulliya,AnticancerRes ,1995,15,295 13 R W Redmond M B Srichai J M Bilitz D D Schlomer and M Kreig Photochem Photobiol 1994,60,348 14 M Krieg J M Bilitz M B Srichai and R W Redmond Blochim Biophys Acta 1994,1199,149 15 C Lee S S Wu and L B Chen Cancer Res ,1995,55,2063 16 K Orth,A Ruck,A StanescuandH G Beger,Lancet 1995,345,519 17 M Motsenbocker H Masuya H Shimazu TMiyawaki Y Ichimori and T Sugawara Photochem Photobiol ,1993,58,648 18 L Strekowski,D -F Hou,R L Wydraand R F Schinazi,J Heterocycl Chem 1%3,30,1693 19 Q A Peng S B Brown J Moan J M Nesland M Wainwright J Griffiths B Dixon J Cruse Sawyer and D Vernon J Photochem Photobiol B 1993,20,63 20 R Y Dodd Trans Clin Biol ,1994,1 18 1 21 M Wilson T Burns J Pratten and G J Pearson J Appl Bacteriol 1995,78,569 22 M Wilson and C Yianni J Med Microbiol 1995,42,62 23 C E Millson M Wilson A J MacRobert and S G Bown J Photochem Photobiol B 1996,32,59 24 M Wilson and N Mia J Oral Path Med ,1993,22,354 25 L Cincotta J W Foley and A H Cincotta Cancer Res ,1993,53,2571 26 L Cincotta J W Foley,T MacEachern,E Lampros and A H Cincotta Cancer Res ,1994,54,1249 27 L Cincotta D Szeto E Lampros T Hasan and A H Cincotta Photochem Photobiol 1996.63,229 28 R C Richmond and J A O’Hara Photochem Photobiol 1993,57 291 29 P Pal H Zeng G Durocher D Girard T Li A K Gupta R Glasson L Blanchard L Gaboury A Balassy C Turmel A Laperriere and L Vrlleneuve Photochern Photobiol ,1996,63 161 30 D Kessel and K Woodburn Br J Cancer 1995,71,306 3 1 A Viola C Hadjur A Jeunet and M Julliard,J Photochem Photobiol B 1996,32,49 32 M Fiedorowicz J R Galindo M Julliard P Mannoni and M Chanon Photochem Photobiol ,1993,58,356 33 S M Burrow,S Guinot,M Wainwright,J J WaringandD A Phoenix Biochem SOC Trans 1995,23 S260 and papers in preparation 34 Y Iwamoto,T Itoyama K Yasuda T Morita T Shimizu T Masuzawa and Y Yanagihara Biol Pharm Bull 1993,16 1244 35 M Tatsuta H Ishi H Yamamura R Yamamoto and S Okuda Oncology 1988,45,35 36 J M Zdolsek APMIS 1993,101 127 37 H Lindauer P Czerney and U W Grummt J Prukt Chem ,1994,336 521

ISSN:0306-0012    

DOI:10.1039/CS9962500351

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Nitrous Acid and Nitrite in the Atmosphere Gerhard Lammel Max-Planck-lnstitut fur Meteorologie Bundesstr. 55 20 146 Hamburg Germany J. Neil Cape Institute of Terrestrial Ecology Bush Estate Penicuik UK EH26 OQB 1 Introduction Why is nitrous acid important as an atmospheric trace gas? It is present in very small quantities typically up to a few parts in lo9by volume (ppbV) and usually makes up only a small fraction of the total amount of gaseous nitrogen oxides present in the atmosphere. One reason for an increasing number of recent studies into the chemistry of HNO is the increasing importance of nitrogen oxides as atmospheric pollutants. Unlike the emissions of sulfur dioxide which have been decreasing markedly in recent years in Europe and North America if not in the developing world the emissions of nitrogen oxides are far from decreasing. In developed countries the major source of nitrogen oxides is vehicle traffic and the steadily increasing number of cars and car journeys more than compensate in many cases for emission reduction measures. On a global scale the importance of NO pollutants (= NO + NO ) is on the rise too because people produce these trace gases directly from a range of activities (e.g.fossil fuel combustion in the energy sector biomass burning in the context of agriculture) or enhance natural sources (e.g. soil NO emission following fertilization with nitrogen). The global significance of the nitrogen oxides in the atmosphere is in determining the production and consumption of the highly reac- tive hydroxyl and hydroperoxyl radicals HO (= OH + HO,). The OH radical is the most important trace species in tropospheric chem- istry and its abundance governs the oxidation and eventual removal of most trace gases from the atmosphere. It is in this respect that HNO shows its importance as a trace gas; on photolysis (A< 390 nm) HNO directly produces OH. This OH source is of greatest importance during the early morning when HNO concentrations may be high after night-time accumulation and when OH production rates from other sources (photolysis of ozone and formaldehyde) are slow. Concentrations of OH radicals later in the day are also influ-enced by HNO,. Apart from its role in tropospheric chemistry nitrous acid is of toxicological relevance. It is suspected to be a pre- cursor of carcinogens (nitrosamines) under atmospheric conditions. ~~ Gerhard Lammel was born in 1960 and studied chemistry in Regensburg and Freiburg Germany where he received a diploma in 1985 then moved to Mainz Germany where he worked for his doctorate (Dr rer. nat.) with the Air Chemistry Department of the Max Planck Institute for Chemistry. From 1988 to 1993 he was afJ- iated to the Karlsriihe Nuclear Research Centre a time period which included a stay as a research fellow with the Lawrence Berkeley Laboratory Berkeley California. Since I993 he has been a research Jcientist at the Max Planck institute for Meteorology at Hamburg. He is a lecturer in Environmental Analytical Chemistry at the University of Hohenheim Stuttgart and works for the German Advisory Council on Global Change WBGU. His main scientijc interests are atmospheric chemistry in clouds and aerosols and inter- disciplinary aspects of global environmental change. Despite this importance nitrous acid is one of the least researched species in inorganic tropospheric chemistry. Why is this so and why do we know so little of its major sources and sinks? The aim of this review is to unravel the puzzle of HNO occurrence in the atmospheric boundary layer close to the earth’s surface as far as present knowledge allows by investigating evidence from both laboratory and field studies. 2 Why do we know so little? The apparent lack of information about HNO in the troposphere is primarily because it is very difficult to make reliable measurements of HNO at ambient concentrations. Conventional chemilumines- cent gas analysers which are used to measure NO (in the chemilu- minescence reaction with 0,) and NO (as NO after reduction) include a signal from HNO as part of the NO signal. The catalytic converters used in these instruments are designed to reduce NO selectively to NO but are likely also to reduce HNO,. In many sam- pling systems HNO may be lost to the walls of sampling lines or inlet filters before it even reaches the catalytic converter because it appears to be a ‘stickier’ molecule than NO or NO,. The potential presence of HNO as part of the NO signal from such analysers is usually ignored. 2.1 Gas-phase Measurement Nitrous acid can be determined in ambient air using differential optical absorption spectroscopy (DOAS) or denuder systems (DS). Spectroscopy as an in situ method is clearly advantageous when dealing with molecules which are reactive on surfaces. Gaseous HNO absorbs light in the near UV with maxima at 354 and 368 nm; its absorption coefficient at these wavelengths is 5.0 X and 4.5 X cm2 per molecule respectively.’ Most HNO concen- tration measurements have been based on earlier cross-section data John Neil Cape studied chemistry at the University of St Andrews Scotland before lecturing in chemistry at the National University of Lesotho (1974-1976). He returned to the Department of Physical Chemistry in Cambridge to study for a PhD using molec- ular dynamics to investigate phase changes in simple systems. He joined the NERC institute ofTerrestrial Ecology as an environmen- tal chemist in 1979 and is now a Principal Scientijic Oflcer in the trace gases research section at ITE’S Edinburgh Research Station. He is also an Honorat-y Lecturer in the Department of Chemistry at the University of Edinburgh. Current interests include a continuing involve- ment with chemical processes at interfaces in particular between leaf surfaces and the atmosphere and with the chem- istry of the troposphere both in the field and in the laboratory. 36 1 which led to an overestimation by ca 5% Much larger discrepan- cies in the absorption cross-section have now been resolved I The first unequivocal detection of HNO in the ambient atmosphere became possible through the development of an optical absorption technique with a rapid-scanning device A slotted disc rotating in the focal plane allows for ca 100 scans of the spectrum per second Rapid scanning of the spectrum is necessary because atmospheric fluctuations with frequencies peaking in the range 1-10 Hz would if resolved dominate spectral features Diode array detectors are coming into use now for HNO measurements Interfering trace species can be identified and corrected for with DOAS using simple spectral deconvolution methods The sensitivity of the method increases with the path length which can be up to ca 10 km The method’s performance is such that under favourable atmospheric conditions a detection limit of ca 0 2 ppbV could be achieved with a time resolution of 15-30 min with an uncertainty in the HNO determination of ca ? 10% The uncertainty of the absorption coef- ficient applied to calculate the concentrations introduces another ca 5 15% uncertainty In a denuder tube (usually made from glass) trace gases are sep- arated from the aerosol in sampled air by diffusion Gases diffuse from a laminar flow regime through the denuder to the walls where they are absorbed by an appropriate coating while aerosols having very much smaller diffusion rates pass through the denuder without being trapped The geometry of the denuder tube and flow rates can be adjusted such that a high (usually close to 100%)and known collection efficiency is achieved Alkaline surfaces mainly sodium carbonate coatings make effective traps for HNO After collec- tion the material is then washed from the surface and the extract is analysed off-line by ion chromatography DS techniques are time- consuming but developments have been made recently in the use of automated denuders No denuder system is specific for HNO A correction for the simultaneous collection of other trace gases such as NO can be made if the interfering gas has a poor collection efficiency With two or more denuder tubes in series nitrite found in the second denuder can be attributed to such an interfering species and the correspond- ing amount would be expected as a positive artefact in the first tube Although various sources of artefact can be identified and avoided undetected reactions of the nitrogen oxides on the surfaces of the DS cannot be completely excluded For instance it has been found that at elevated levels of SO an interference of NO can cause sig- nificant positive artefacts on Na,CO coatings This might be the reason why high HNO abundances during day-time obviously not in agreement with the species’ photochemical properties have been reported on the basis of DS measurements In principle very low concentrations should be measurable by this method by using long sampling times However because some interferences tend to become more influential at long times (eg the slow diffusion of small particles to the denuder walls) the detection limit for mea- suring HNO by DS is limited too However DS can be applied under conditions of high humidity when higher extinction prevents absorption measurements Recently a modified chemiluminescence technique has been developed,6in which HNO is absorbed into solution using a diffu- sion scrubber then selectively reduced by ascorbic acid to NO which can be determined using the conventional chemilumines- cence technique The use of a diffusion scrubber is claimed to avoid interferences from NO This technique however is limited by the detection limits of the NO analyser employed and by the need to measure the HNO concentration in the presence of NO as the dif- ference of two large numbers (cNo+ cHNO2)and (cNo)The use of a diffusion scrubber followed by the direct quantification of the nitrite ions produced in solution by colorimetry or ion chromatog- raphy has also been demonstrated but there are few published applications of this technique Although several methods for detect- ing HNO in ambient air have been developed it is not yet possible to purchase a ‘Nitrous Acid Analyser’ and most of the measure- ments to date have been made using instrumentation developed in a research laboratory As a result there are few instances of long- term application of such instruments to characterising the HNO concentration at a site over months or years CHEMICAL SOCIETY REVIEWS 1996 2.2 Aqueous-phase Measurement Dissolved HNO and nitrite Nil1,associated with particles or fog and cloud droplets is determined by wet chemical or ion chro- matographic methods A detection limit of a few ng m can be achieved for particles by integrating over many hours ( 12-24 h) and ignoring the size distribution An additional risk with long Sam pling periods is the possibility that aqueous samples might be subject to rapid oxidation Therefore on-line analysis or stabiliza- tion through a derivatization reaction is strongly recommended On- line measurement of nitrite in cloud water has been achieved? with detection limits around 10 nmol dm 3 Observations of Nitrous Acid in the Troposphere 3.1 Field Measurements of Gaseous Nitrous Acid Since the first unequivocal detection of HNO in the atmosphere? this species has been observed together with NOxin the night time atmosphere at both urban and rural sites and also (at distinctly lower concentrations) during day-time Most of the measurements have been made during winter and spring although there are now a few sets of year-round data Night-time HNO concentrations up to several ppbV have been observed Concentrations in rural areas are not necessarily lower than in regions with strong NO sources (Figure 1) Rural concentrations over a two year period in Lower Saxonyio averaged 0 6 ppbV but with daily maxima in winter up to 8 ppbV HNO has also been measured inside buildings and vehicles,S I where very much larger concentrations may be found up to 40 ppbV After sunset when photolysis ceases concentrations of HNO gradually increase In many cases the HNO concentration sta- bilises (more precisely the HN0,-to-NO ratio stabilises) and remains constant throughout the night until after sunrise it starts to decay (Figure 2) During a measurement campaignI2 in the German city of Mainz for instance 38 out of 43 night-time periods of several hours were identified with a constant (insignif- icantly varying) or a continuously increasing HN0,-to-NO ratio whereas this ratio significantly decreased in only five out of 43 nights analysed The observation of a constant HN0,-to- NO ratio can be explained in one of two ways It is consistent with (a)a constant reaction time relative to an upwind source as air is advected to the measuring site or (b)the balancing of the rate of formation of HNO by the rate of removal of HNO in a stagnant or well mixed air mass Source strengths at the site and upwind together with transport and dispersion processes need to be considered explicitly In case (a)pollutant concentrations at the site are dominated by advection rather than local emissions and a horizontally homogeneous atmosphere is implied This can be classified as a ‘receptor site situation’ where chemical species observed have the same fixed reaction time as given by the trans- port time from the emission site The closer the primary pollu- tants sources the lower are the secondary pollutant levels Dispersion processes have diluted primary pollutant concentra- tions since emission Case (b)is the situation close to the emis- sion sources and is comparable to a smog chamber during filling Pollutants and their secondary products accumulate through direct emission or chemical transformation processes and inefficient mixing with outside air This situation is common in urban areas where primary emissions occur continuously and mixing with cleaner air is suppressed at low wind speeds below a low night-time inversion layer In fact a broad distribution of reaction times (with respect to the time of emission) is then addressed because measuring techniques integrate over a certain time period (minutes to hours) In Figure 2 the time variation of pollutants including HNO is shown on a night when as wind speed dropped a shift from case (a)to case (b)occurred The observation of decreasing HN0,-to-NO ratios can be caused by (a)the mixing of fresh NO emissions into the plume from the upwind source (receptor site situation) or (b)from a particularly NITROUS ACID AND NITRITE IN THE ATMOSPHERE-G. LAMMEL AND J. N. CAPE ppbV HONO 0.1 1 10 1000.01 Los Angeles area Los Angeles area Los Angeles area Los Angeles area Lubbock Texas Lubbock Texas Chicago Cologne Malnz Gothenburg Gothenburg Milan Milan MtMitchell,NC,USA San Gabriel Mtns.,CA day4 San Gabriel Mtns.,CA night4 Slerra Nevada CA,USA Claremont CA,USA Riverside CA USA Jullch,Germany Jullch,Germany Braunschwelg,Ger. summer Braunschwelg,Ger. wlnter SE England SE England SE England near Rome Italy near Rome Italy Ispra N.ltaly Chro Swltzerland Alert Canadian Arctic I SW desert Utah,USA savanna Veneuela Loop Head Ireland Figure 1 Summary of published HNO measurements in urban rural and remote areas shown as the average (H)and range. Data from some sites are restricted to a few hours while at other sites the data represent continuous measurements over a year. The diamonds (0)show the average ratio of HNO to NO for sites where NO concentrations were measured simultaneously. A detailed reference list is available from the authors. effective HNO decay reaction (reverse reaction producing NO,) in a stagnant or well-mixed air mass. So far indications for the existence of a reverse reaction are sparse. In the Mainz data set concentrations of primary pollutants (HCHO-to-NO ratio) indi- cated that the decrease in the ratio was caused by an increase in primary emissions of NO,. In two nights of constant HN0,-to- NO ratios the concentrations of the primary pollutants indi- cated the advection of increasingly older emissions to the site suggesting the influence of a mechanism which removed HNO from the air mass. 3.2 Correlations with other Trace Gases Concentrations of HNO are often correlated with those of NO and NO,. This is partly a causal link because NO and NO are source gases for the formation of HNO or may have a common source. However all gases in the atmosphere are subject to coincident con- centration changes as the depth of the boundary layer changes (for example under a temperature inversion in winter) or as wind direc- tions change to bring a different mixture of trace gases past a fixed measurement point. This strong dependence on meteorological con- ditions has bedevilled attempts to understand the sources and sinks 364 I 1987-05-05/06' haze wind N (1023)m/s; after 0O:OO c 3 m/s 1.5 T = (10.7k 1.2)'C +HNO IppbVl*502 I100 ppbvI 1,-4 HCHO [ 10ppbv 1 H) 22:OO 0O:OO 02:OO 04:OO 06~00 OBiOO 10~00 5.5.1987 6.5.1987 I MESZI sunset sunrise Figure 2 Measured HNO concentrations at Mainz in a heavily industrial- ized area. Until midnight (OO:OO) strong winds prevailed and pollutant concentrations at the site were probably dominated by advection rather than local emission. Source strengths had decreased and dispersion processes had diluted primary pollutant concentrations since emission. Nitrous acid was formed after sunset and disappeared during the morning. Around midnight as wind speed dropped HNO accumulated and the HN0,-to-NO ratio increased significantly. of HNO on the basis of field measurements alone. Nevertheless considerable effort has been made to understand the processes in the atmosphere which lead to HNO formation because the amounts of HNO observed are often an order of magnitude larger than would be expected on the basis of reaction rates determined in the labora- tory. This is a hotly debated problem in tropospheric chemistry and will be discussed below. 33 Observationsof Nitrite and Possible Chemical Formation In comparison to other ions only a few N"' measurements in en- vironmental samples have been reported so far mainly because con- centrations of N"' are low and nitrite ions are easily oxidised. 3.3.1 Cloud and Fog Water In the anthropogenically influenced atmosphere cloud and fog water is usually more acidic than the natural composition of the air would suggest (which would be ca. pH 5.6 from carbonic acid). Acidity is produced by the oxidation of anthropogenic non-metal compounds in the atmosphere primarily SO and NO or in some cases by the solution of acidic gases such as HCI. Pollutant con- centrations are particularly enriched in radiation fogs as a result of the small size and long lifetime of the droplets. According to the dis- sociation equilibrium of HNO (pK = 3.5) and its modest solubil- ity [Henry coefficient H = 49 mol dm- atm-I at 298 K; 1 arm = 101 kPa)] only very small concentrations of N"' can be achieved in acidic fog or cloud water droplets. The situation is different though under conditions of low acidity or neutrality. These tendencies are reflected in the measured Nil' cloud and fog water concentrations compiled in Figure 3. A partial pressure of 1 ppbV of HNO would lead to a concentration of N"' < 0.1 pmol dm-3 at pH = 3 but to 15 pmol dm-3 at pH = 6. Unfortunately so far none of the studies which addressed cloud or fog water N"I included co-located gas- phase HNO measurements to test Henry's Law equilibria in the field. 3.3.2 Particulate Matter Particulate nitrite concentrations observed are in the lower ng m-3 range (pptV parts in 1OI2 by volume) reaching up to hundreds of ng rn- (>0.1 ppbV) in the urban area of Milan (Figure 3). When measured together with gas-phase HNO on a short-term basis (up to a few hours) no clear relationship between the two was detected.13 This is no surprise because particulate nitrite through its dissociation equilibrium is in a pH-dependent phase equilibrium CHEMICAL SOCIETY REVIEWS 1996 Figure 3 Summary of published measurements of nitrite in particles (expressed as equivalent mixing ratios ppbV) or dissolved in cloud and fog water (as concentrations in pmol dm-'). The range and average (m) concentrations are shown. For particles simultaneous gas-phase concen- trations of HNO (ppbV). where available are shown as filled diamonds (+). For cloud and fog samples average NO concentrations during sam- pling (ppbV) are shown by open diamonds (0).A detailed reference list is available from the authors. with HNO,. The atmospheric aerosol on these occasions was a source of gas-phase HNO because particulate matter Nil1 concen- trations were in excess with respect to the phase equilibrium. The phase equilibrium of particulate nitrite can be seen as analogous to the situation in cloud and fog water because an aqueous layer covers atmospheric particles under ambient humidities. In the Arctic where nitrite concentrations were measured in particles which were approximately 70% (mlm) H,SO the atmospheric aerosol also appeared to be a source of HNO with equilibrium between gas and particulate phases.8 4 Atmospheric Chemistry of Nitrous Acid Nitrous acid and nitrite are short-lived intermediates of reactive oxi- dised nitrogen in the atmosphere (collectively referred to as NO,). NO enters the troposphere in the form of NO mainly as NO. The main fate of NO is to combine with the OH radical (during daytime) or to be oxidized to the nitrate radical NO by ozone leading to nitric acid HNO or particulate or cloud- and fog-water nitrate NV. NO is removed from the atmosphere mostly in the form of HNO or particulate nitrate ions through wet and dry deposition and by these processes contributes considerably to acidification and eutrophication of ecosystems. The occurrence of HNO or nitrite though of great significance for atmospheric chemistry does not have any significant impact on the nitrogen balance. NITROUS ACID AND NITRITE IN THE ATMOSPHERE-G LAMMEL AND J N CAPE 4.1 Sources of Nitrous Acid Combustion processes are a direct but small source of HNO Emissions from vehicle exhaust pipes have a mole ratio HNO,/NO of ca 0 15%,and rapid reactions subsequent to emission could lead to ratios up to 0 5% 14 Measurements during the burning season in the tropical savannah suggested that HNO is also a by-product of biomass burning I5 In both rural and urban areas however vehicu- lar emission of HNO cannot explain the high abundances the HNO,/NO ratio typically reaches a few percent (Figure 1) Production of HNO is far more efficient than expected from homo- geneous gas-phase reactions of the nitrogen oxides which overnight could account for at most 0 5% HNO,/NO There IS clear evidence that the formation of HNO in rural and urban environ- ments is governed by heterogeneous reactions The initial build-up after sunset should reflect the chemical source Formation rates dcHNOjdt up to 1 3 X ppbV s-I and conversion rates (dcHN02/dt)/cNOz up to 4 X S-I have been reported from a rural and an urban site in Germany l6 dcHNojdt was roughly proportional to cNO2,which suggests first-order chem- ical kinetics 4 I I Possible Pathways from Laboratory Measurements What are the possible chemical pathways' The chemistry of nitro- gen oxides and nitrogen oxyacids comprises a wide variety of reac- tion paths Only part of it applies to the concentration range relevant to the atmosphere Reactions R1 to R4 (below) have been suggested to explain the formation of NIIi in the atmosphere As OH levels are very low during the night channel R3 can only be significant during day t ime (Rl/R- 1) 2 NO + H,O = HNO + HNO (R2/R-2) NO + NO + H,O = 2 HNO (R3/R-3) NO + OH = HNO (R4) NO + Red,? = NO + Oxads Early smog chamber studies dedicated to HNO were stimulated by the detection of an interfering radical source It was then shown by different laboratories that gas-phase HNO formation from NO in humid air on a wide variety of surfaces (glass borosilicate glass quartz fluorinated polymers sodium halides oxalic acid) proceeds with a stoichiometry similar to the aqueous phase namely via R1 The rate-determining step however is first order with respect to both NO and H,O concentrations Is NV remains strongly adsorbed to the surface The same applies to HNO occurrence indoors I l9 Isotope studiesi7 suggested a mechanism which includes the dissociation of NO but could not completely clarify the mechanisms R1 is particularly fast on cleansed glass and on metal surfaces In smog chambers metal surfaces are always present in the form of end plates and mirrors of optical components Thus the apparent formation rate in smog chambers has to be regarded as a superposition of different surface activities and a simple increase in available surface area usually will not lead to an equal increase in the apparent rate of a surface reaction It is nonetheless clear that the reaction rate is proportional to the surface area The rate constants on passivated surfaces of different materi- als are in fairly good agreement with a range of k = (1 5 -6) X 10 22 cm4s-I per molecule (for dcHNO!dt = kI S V-I c~ at room temperature) l7 l8 Rate constants measured in a dark fur-nished mobile laboratory fall in the same range I9 On alkaline sur- faces like Na,CO source reactions R 1 and R2 both contribute and R4 accounts for nitrite formation on denuders coated with guaiacol (an aromatic compound in a reduced state) 2o 4 I 2 Possible Chemical Pathwats fiom Field Measurements From field investigations it is clear that NO is a precursor of HNO at both urban and rural sites (suggesting R 1 or R2 or both) In many cases R2 has been rejected because measured NO concentrations were too small2' or indirectly through the detection of high ozone concentrations which do not allow for NO to be present due to the very efficient 'titration' of NO by 0 In several field studies in urban areas a role for NO (R2) has been claimed and a mechanism has been proposed involving formation of N,O as an intermedi- ate 22 Justification for this route has been sought on the basis of strong correlations between NO and HNO concentrations NO is the predominant primary emitted NO molecule and NO is formed from NO within minutes as long as 0 is available At a site close to sources of NO however where primary and secondary pollu- tants may accumulate (for example in a stagnant air mass) coinci- dental occurrences of NO and HNO are to be expected independent of any role of NO as a precursor of HNO (as discussed above) Thus neither in the laboratory nor in the field 1s there any conclu- sive evidence of a significant contribution of R2 to HNO forma- tion With respect to water vapour no simple relation between HNO build-up and humidity has been found in the ambient atmos- phere 4 I 3 Heterogeneous Reactions Two types of surface are present in the ambient atmosphere namely the ground surfaces (the bare soil biological surfaces walls of buildings paved roads) and the surface of airborne particles Evidence for the involvement of both surface types has been found A strong coincidence of HNO and radon (which has sources exclu- sively at the ground) concentrations in central Milan13h suggested a ground-based source of HNO The surfaces of buildings and streets are substrate candidates NO is known to be very efficiently absorbed by cement and building stone both dry and moist 23 The ground surface may act as both source and sink for HNO over grassland in SE England an upward flux of 5-25 ng HNO m s I (which would correspond to tenths of a ppbV HNO h I under the influence of a low night-time boundary layer) was observed when NO concentrations exceeded 10 ppbV while at lower concentra- tions a net downward flux occurred24 (dry deposition velocity vHNO = 0 2-1 7 cm s-l) The importance of aerosol surfaces was suggested by the obser- vation that production of HNO was particularly effective under hazy conditions l6 2s A correlation between HNO mixing ratios and particulate matter surface area (as estimated by a nephelomet- ric method) was observed at Julich Germany 26 A clear correlation was also seen in rural air in northern Italy where co-variation with the boundary layer depth was minimised 27 Furthermore whenever aerosol has been analysed for nitrite it has appeared that atmos pheric aerosol was a source of gaseous HNO l3 It is unclear however whether the aerosol surface source was the only or the dominant one The nitrate radical NO and its adduct N,O are interesting intermediates in the atmospheric chemistry of nitrogen oxides Laboratory and field experiments demonstrated that N,05 forms nitrate on sea-salts urban particulate matter and on fly ash In one study however when applying ppmV (parts in lo6by volume) con- centrations nitrite rather than nitrate formation was suggested 28 By this pathway HNO could be formed on the surface of sea-salt aerosol in remote regions where the lack of NO would permit the formation of NO and thus N,O The relevance of this finding for ambient NOx concentrations is not clear particularly in the light of more recent evidence which suggests that negligible HNO is pro duced by this pathway 29 The heterogeneous decomposition of per-~~c ~~nitric acid HO,NO,~ (PNA) or peroxyacetyl nitrate MeC(O)OONO (PAN) would also lead to nitrite in the aqueous phase30 (as a function of pH in the case of PNA) and then HNO in the gas-phase Neither source is thought to be of any significance except perhaps at high altitudes or at high latitudes during winter- time In the laboratory reactions with model aerosols such as carbon and metal oxides have been studied Oxides of some metals appar- ently enhance HNO formation from NO in humid air while others do not and common characteristics are not obvious 3l When in contact with activated carbon and humid air NO produces NO Nili and NV When NO was bubbled through a slurry of activated carbon much more nitrite than nitrate was produced in clear devi ation from known stoichiometries 32 NO under reduction to Nil1 enhances greatly the oxidation of SO on freshly generated soot and combustion aerosols 33 The importance of R4 in the tropos- phere i e direct reduction of adsorbed NO is speculative to date NO is a strong oxidant (potential E = 1 065 V) R4 requires the presence of reductive surfaces which appears unlikely for ground surfaces which are continuously exposed to an oxidative environ- ment For aerosol surfaces the reduction of ppbV levels of NO for c,,,,~ = 100 pg m-3 would mean the presence of reductants in the particulate phase in the percent range (by mass) which also appears unlikely However the reductant could be fed from the gas-pha~e~~(SO,) or recycled in the polluted aqueous phase (like Mn+ recycled from the reduction of M(n+l)+by dissolved SO M = Fe or other) The production of NO from soil and stone samples upon NO uptake has been related to the presence of oxidizable metal ions 23h In areas with intensive agriculture ammonia concentrations are often sufficient to completely neutralize airborne acids Because the solubility of both NO and SO is strongly favoured by high pH the reduction of dissolved NO by dissolved SO may become impor- tant SIv however would be prevented from oxidation in the pres- ence of formaldehyde (HCHO) because it forms the adduct hydroxymethane sulfonate which is particularly stable in the pH range 3-6 Figure 4 shows the results of a model of Nili concen-trations in aqueous solution representative for fog and cloud water as a function of both gas-phase pollutants and pH under both low and high HCHO and HNO concentrations In the absence of gas- phase HNO micromolar levels of Nil1can be expected from NO reactions in the presence of SO but only if SO is available [compare Figure 4(6) and 4(c)] Production of Nil1in the aqueous phase at low HCHO in the absence of gaseous HNO [Figure 4(6)] gives concentrations within a factor of two of NII1concentrations observed in the presence of gaseous HNO at pH 6 [Figure 4(6)] In the presence of gas phase HNO production of NII1by reactions involving SO (in combination with HCHO) is less important leading to differences in N111 of a factor of three at most as far as the parameter variation of Figure 4 is concerned This model sug- gests that part of the Nili observations in neutral or almost neutral atmospheric aqueous samples could be explained without HNO originating from the gas-phase This was the case for the Stuttgart fog events (Figure 3) 3s The observation of significant concentrations of nitrite in cloud water (Figure 5) led to the suggestion that nitrite was formed in or on cloud droplets and that this was the most likely pathway to explain the conversion of NO into other oxidised nitrogen com- pounds (measured as the sum NO,) during the passage of air through hill cap clouds in Northern England 36 4.2 Sinks for Nitrous Acid Photolysis (backward reaction R-3) is the dominant sink during day-time although photolysis of HNO is around four times slower than NO photolysis and is on a timescale of tens of minutes So far indications for the existence of a night-time removal reaction in the gas phase are sparse The frequently observed achievement of a constant HNO,/NO ratio might be caused by transformation as well as transport processes or both (as discussed above) Dry depo- sition and removal at the ground can also deplete concentrations close to the earth's surface In smog chambers an equilibrium of heterogeneous HNO formation and decay reactions is established after some time NO appears as a product of this removal reaction in the gas-phase (wall materials borosilicate glass quartz poly- ethylene fluorinated polymer surfaces) l7Is This could be attrib- uted to R-2 which is known to proceed at least partially heterogeneously Its rate is influenced by water vapour concentra- tion and surface type There is no evidence for R5 (below) taking place unless at high cHNO2and on activated carbon surfaces At higher humidi- ties NO production is enhanced indicating a non-trivial mech- anism When in contact with acid surfaces (dry quartz) or when bubbled through acidic solutions HNO reacts in several steps to form nitrous oxide N,O as well as nitrate 37 These findings suggest that with strong acids such as H,SO HNO acts as a base (R6) CHEMICAL SOCIETY REVIEWS 1996 (R5) HNO + NO -HNO + NO (R6a) HNO + H,SO c' NOHSO + H,O (R6b) 2 NO+HSO + 2 HNO + H,O -2 HNO + 2 H,SO + N,O Stone surfaces as well as acting as potential sources of HNO from the heterogeneous reaction of NO with water may also act as a sink Colonies of nitrifying bacteria are frequently found on the surface of stone monuments They produce NO from NH,' deposited on these surfaces with NO; as an intermediate 3s The presence of gaseous SO suppresses NO production which sug- gests that N1I1formed at surfaces might also have a chemical sink besides evaporation and contribution to gas phase HNO 5 Can Field and Laboratory Measurements be Reconciled? How can we translate laboratory measurements of both homoge- neous and heterogeneous reactions into the field? During the night boundary layer heights of 50-100 m can be assumed For a column of air with a base area of 1 m2 the ground surfaces provide ca 10 m2 from structures (a factor of 6-14 for vegetation surfaces and ca 10 for urban surfaces such as buildings and roads) and the aerosol 0 005-1 m2 (S/V = 10 m for rural and urban aerosols times a factor of up to ten to account for particle porosity) leading to a specific surface to volume ratio (S/V)between 0 1 and 1 2 m I Applying the rate for passivated surfaces found in labora-tory experiments k = (1 5-6) X lop2,cm4s I per molecule an apparent conversion rate (dc,N,jdt)/cN,2 of (0 01 -3) X 10 6 s-1 would be expected for the relevant temperatures and humidities Even the upper estimate cannot explain the highest observed con- version rates of 4 X 1 0-6 s I Moreover these apparent conversion rates are only lower bounds for the effective NO to Nili conversion rate in the field because under the influence of the cooling after sunset HNO desorption from surfaces to the gas phase will be sup- pressed Atmospheric surfaces can be of a reactivity comparable to that of a freshly cleaned smog chamber wall (borosilicate glass) I7h *Sh The aerosol surface if the site of HNO production must be much more efficient than laboratory surfaces studied so far At present one can only speculate about the reasons Real surfaces such as particles urban surfaces and the bare soil may be more active due to the pres- ence of ions capable of contributing to charge-transfer type reac- tions transition metal concentrations in ambient aerosols are considerable Ambient aerosols rich in carbon although less reac- tive than model substances like activated carbon are prime candi- dates when searching for active surfaces they provide a high specific surface area and carry a wide variety of functional groups and adsorbed material When newly formed such particles are hydrophobic but may eventually be incorporated into fog and cloud water The wetting of surfaces could mean that active sites are blocked and pores become inaccessible resulting in a reduced spe cific surface area Wetting of particles occurs as a function of humidity because they take up liquid water above the humidity of deliquescence of the predominant salt components The growth curve of ambient aerosol rises steeply above 70-75% relative humidity (r h ) Under these conditions one might expect HNO production to be inhibited and this idea is supported by observa- tions that only in very few cases could significant HNO formation be observed at r h > 70-75% and then only when advection was from a sector with a large fraction of urban surfaces 39 Nitrite can also be formed on the surface of particles and water droplets by the same reactions (R1 and R2) which give gaseous HNO It is here that the greatest interest and uncertainty lies because this heterogeneous route may yet prove to be one of the most important in reconciling laboratory and field measurements The reaction of NO with bulk liquid water is very slow at ambient concentrations because it is a second-order reaction 40 Recent mea- surements suggest however that this reaction may not be so simple on finely divided water droplets and the possibility of first-order behaviour has been raised Experimental determinations of the reac- tion rate in simulated clouds are not easy and some controversy still 50 50 45 45 40 40 35 35 n 30 z -302 Q.0. n v v 0" 25 25 z z P 20 P 20 15 15 10 10 5 5 I 3 I I 4 I I 5 I 4 5 6 PH PH (d) 50 45 40 35 ns= Q0 v p 25 P P 2 20- 14I 15-16 5- 3 4 5 6 PH PH Figure 4 Modelled N1" concentrations in aqueous solutions representative of fog and cloud water as a function of both NO and pH Calculations were made with a two-phase box model of chemistry and interphase transport 3s For the system under study mass transport limitations are negligible (assumed droplet size 5-10 km) Total solute concentration is ca 2 5 mmol dm 3 reaction time is 1 h Aqueous phase chemical kinetics include the reaction of S'" species with dissolved NO (A(*) = 1 3 X lo7dm mol-I s I for weak acid and neutral media) which is the main source of N"' in the absence of gas-phase HNO Only cases with equal concentrations of the gas phase pollutants NO and NO are considered Initial concentrations (at t = 0) are (a)HNO = 0 SO = 0 (b)HNO = 0 SO = NO HCHO = 0 5 ppbV (c)HNO = 0 SO = NO HCHO = 10ppbV (4HNO,/NO = 0 05,SO = 0 368 [H+]/mmoldm 4 0.03 I I I I I i 0.6 -;. I. I w.w 1400 1600 1800 2000 2200 2400 hme (GMT) calculated HN02 (ppbV) 0.25 I 1 I 1 I 0.20 t L0.15 t 0.10 ?. 1400 1600 1800 2000 2200 2400 time (GMT) Figure 5 Measurement of cloud water N"' concentrations in real time on 7 November 1991,on the summit of Great Dun Fell Cumbria England w Direct measurements of N"' and pH are used to infer gas phase concen trations of HN02 on the assumption of Henry's Law equilibrium The temporal pattern of gas phase HNO concentrations correlates with con centrations of NO reflecting the influence of different air masses either as sources of gas phase HNO before entering the cloud or as precursors to HNO formation on droplet surfaces The dashed lines show interpola tions between measurements GMT=Greenwich mean time exists on their interpretation 41 Other indications of possible het- erogeneous reactions at water surfaces have come from measure- ments of the accommodation coefficient of gaseous NO on a liquid waterjet with measured uptake of NO being much larger than cal- culated 42 Until the mechanisms of these heterogeneous reactions are understood the modelling of urban photochemistry which is dependent on the correct formulation of HNO chemistry may have to resort to empirical expressions As an example of this approach the co-variation of NO NO HNO in the Los Angeles area has been used to derive a pseudo-homogeneous reaction rate for R2 and apply it in a model of HNO production 22 Whether this empirical expression will be useful in other regions with very different mixes of precursor gases remains to be seen 6 Conclusions Nitrous acid is an important trace gas for understanding tropos- pheric chemistry largely because of its role as a source of the OH radical early in the day Techniques are only now becoming CHEMICAL SOCIETY REVIEWS 1996 available to make routine measurements possible although they are labour intensive and as such data become available a better picture of the overall importance of HNO will be gained Despite a long- standing awareness of the heterogeneous reactions which link NOx and HNO there are still few experimental data from laboratory studies which are directly applicable to ambient conditions of con-centration temperature pressure humidity and active surfaces Moreover there appear to be large gaps in our understanding of the underlying mechanisms of the reaction of NO and liquid water Until these gaps are filled we have to accept that laboratory data cannot provide the numbers necessary for use in chemical model- ling of HNO production and loss in the troposphere We will have to rely on parametrisations based on field measurements or take up the challenge and characterise these heterogeneous reactions in the laboratory under realistic atmospheric conditions References 1 A Bongartz J Kames U Schurath C George P Mirabel and J L Ponche J Amos Chem 1994,18 149 and references therein 2 D Perner and U Platt Geophvs Res Lett 1979,6,917 3 A Sjodin and M Ferm,Atmos Environ 1985,19,985 4 P K Simon and P K Dasgupta Environ Sci Technol 1995,29 1534. J Slanina and G P Wyers Fresenius J Anal Chem 1994,350,467 5 A Febo and C Perrino,Atmos Environ 1991,25A 1055 6 Y Kanda and M Taira Anal Chem 1990,62,2084 7 Z VeEeia and P K Dasgupta Environ Sci Techno1 1991,25,255 8 S M Li J Geophvs Res 1994,99,25479 9 J N Cape,K J Hargreaves,R Storeton West,D Fowler,R N Colvile T W Choularton and M W Gallagher Amos Environ 1992,26A 2301 10 R Zimmerling U Dammgen A Kusters L Grunhage and H J Jager Environ Pollut 1996,91 139 11 M Brauer P B Ryan H H Suh P Koutrakis and J D Spengler Env Sci Technol 1990,24,152I A Febo and C Perrino Atmos Environ 1995,29.345 12 G Lammel PhD Thesis University of Mainz Germany. 1988 13 (a) G Lammel and D Perner J Aerosol Sci 1988 19 1199 (b) I Allegrini A Febo and C Perrino In Phvsico-chemical behaviour of atmospheric pollutants ed G Restelli and G Angeletti European Commission Report No EUR 1560911,pp 293-299,1994 14 D Perner C Kessler and U Platt in Moniroring of gaseous pollutants bv tunable diode faserr ed R Grisar H Preier G Schmidtke and G Restelli) Reidel Dordrecht 1987,pp 116- 119 15 A Rondon and E Sanhueza Tellus 1989,41B 474 16 C Kessler. PhD Thesis University of Koln Germany 1984 17 (a)F Sakamaki S Hatakeyama and H Akimoto. Int J Chern Krner 1983 15 1013. (b) R Svensson E Ljungstrom and 0 Lindqvist. Atrnos Environ 1987,21 1529 18 J N P1tts.E Sanhueza,R Atkinson W P L Carter,A M Winer,G W Harris and C N Plum Int J Chetn Kinet 1984 16 919 M E Jenkin. R A Cox and D J Williams Atmos Environ 1988.22,487 19 J N Pitts TJ Wallington. H W Biermann and A M Winer Atmos Environ 1985,19,763 20 F de Santis A Febo and C Perrino Ann Chim (Rome) 1987,763 21 G W Harris W P L Carter A M Winer J N Pitts U Platt and D Perner Environ Sci Technol 1982,16,414,A M N Kitto and R M Harrison Atrnnr Environ 1992,26A 235 22 J G Calvert G Yarwood and A M Dunker Res Chern lntermed 1994,20,463 23 (a)H S Judeikis S Siege]. T B Stewart H R Hedgpeth and A G Wren in 'NitrogeneouJ Air Pollutants -Chemical and Biological ltnplicatioris ed D Grosjean Ann Arbor Science MI. 1979,pp 83 -109,(b)M Baumgartner E Bock and R Conrad. Chetnosphere 1992 24 1943 24 R M Harrison and A M N Kitto. Atmos Environ 1994,28 1089 25 (a) G Lammel D Perner and P Warneck in Phvsico-chernical Behaviour of Attnospheric Pollutants ed G Restelli G Angeletti Kluwer Dordrecht 1990,pp 469-476 (b)J Notholt J Hjcarth and F Raes Atmor Environ 1992,26A 21 1 26 C Kessler and U Platt in Phvsico chemical Behaviour of Atmospheric Pollutants ed B Versino and G Angeletti Reidel Dordrecht 1984,pp 412-422 27 M D Andres Hernandez J Notholt J HjOrth and 0 Schrems Atmos Environ 1996,30. 175 28 W Junkermann and TIbusuki Atrnos Environ 1992,26A,3099 29 R Vogt and B J Finlayson Pitts Geophys Res Lett 1994,21,2291 30 G Lammel D Perner and P Warneck J Phys Chetn 1990.94,6141 NITROUS ACID AND NITRITE IN THE ATMOSPHERE-G LAMMEL AND J N CAPE T Zhu. G Yarwood J Chen and H Niki. Environ Sci Technol 1993 27,982 D Grosjean K Fung J Collins J Harrison and E Brertung Anal Chem 1984,56,569 31 H M ten Brink J A Bontje H Spoelstra and J F van de Vate Studies Environ SCI 1978 1,239 32 L A Gundel N S Guyot-Sionnest and T Novakov Aerosol Sci Technol 1989,10,343 33 L G Britton and A G Clarke Atmos Environ 1980 14,829. W R Cofer D R Schryer and R S Rogowski Atmos Environ 1981 15 1281 34 F de Santis and I Allegrini Amos Environ 1992,26A 3061 35 G Lammel in Air Pollution IV ed H Power B Caussade and C A Brebbia Computational Mechanics Publishers ,Southampton in press 1996 36 R N Colvile T W Choularton K N Bower M W Gallagher J N Cape D Fowler K J Hargreaves. G J Dollard T J Davies S A Penkett R A Burgess and B J Bandy Proc EUROTRAC Svtnp I444 ed P M Borrell et a1 SPB Academic Publishers The Hague 1994,pp I1 14-11 18 37 K H Becker J Kleffmann R Kurtenbach and P Wiesen Fcrrcrdai Discuss Chetn Soc 1995,100. 121 38 E Bock and W Sand J Appl Bacteriol 1993,74,503 39 Lammel and Perner unpublished 40 Y N Lee and S E Schwartz J Phvs Chern 198 I 85.840 J N Cape R L Storeton West S F Devine R N Beatty and A Murdoch.Attnor Environ 1993,27,2613 41 A Bambauer B Brantner M Paige and T Novakov Ammo Environ 1994,223,3225,S E Schwartzand Y N Lee,Attnoh Environ 1995,29. 2557 T Novakov Atmos Environ ,1995,29,2559 42 S Mertes and A Wahner J Phbs Chem 1995,99 14000

ISSN:0306-0012    

DOI:10.1039/CS9962500361

出版商:RSC    

年代:1996

数据来源: RSC