Instruments for LIF are varied and frequently tailored to their application. All methods have several things in common: a laser source, a compartment for enclosing the sample, and a detector. Excitation/detection schemes are also varied and experiments may employ resonance, double resonance, two color, off resonance, and other means of excitation. Thought must be given to the sensitivity of the technique, but cost, size, and technical complexity also factor into design decisions. To achieve the lowest limit of detection (LOD), the excitation and detection frequencies that produce the most intense emission are typically chosen. However, this may not always be the best method. Spectral interferences can hamper detection at some emission lines and not others. In other instances, LOD is not an issue, and linear dynamic range or measurement accuracy is more desirable. This is why experimental setups vary, not only due to the substrate, but the overall goal of the technique.

Laser excited atomic fluorescence spectrometry (LEAFS) is a technique in which volatilized atoms are characterized by the emission of radiation from an excited state that is induced by a laser source [20]. The distinction between LEAFS and atomic fluorescence spectrometry (AFS) is that the atomic excited states are produced from a coherent laser source with a narrow frequency profile, rather than incoherent, broad-band light source. Many papers use LEAFS and LIF interchangeably when referring to atomic systems, since LEAFS is a type of LIF. In this work, LEAFS will be used when referring to atomic species.

The advantages of LEAFS stem from the high degree of sensitivity and selectivity resulting from the selection of excitation frequency and fluorescence frequency [21-23]. The extreme selectivity of the technique makes multi-element measurements cumbersome, though there are recent advances in “multidimensional” detectors [24]. Additionally, a degree of technical complexity is inherent in the system. A general schematic for LEAFS is depicted in Figure 1, where laser irradiation should be tunable between 180 and 800 nm [25]. The laser pulse passes through a cell that contains and/or produces gas phase atomic elements. The resulting fluorescence is carefully selected and detected. Unlike typical LIF experiments, laser intensity in LEAFS is typically set slightly above the limit of saturation in order to reduce the absorption of fluoresced radiation by unexcited species and to minimize the effects of fluctuations in laser intensity [23]. The deleterious effects of non-linear phenomena are minimized, while minimizing self-absorption produces a typical linear dynamic range of 5–7 orders of magnitude.

Unlike atomic emission spectrometry (AES) and AFS, atoms do not fluoresce unless they are excited by laser radiation. A comparison with other laser-based atomic spectroscopy techniques has been made by Sjöström [22]. Since this review is not focused on instrumentation or theory, readers should consult other papers on these subjects [20,21,23,26-29].

The elements investigated by LEAFS must be in the gas phase. If the sample has insufficient vapor pressure, energy must be imparted to the sample to volatilize the element. Volatilization techniques include simple effusive cells, flames [26,30-32], electrothermal atomizers (ETA) [21,33,34], glow discharge cells, inductively coupled plasma (ICP) [20,35,36], and laser ablated (LA) plumes. In general, the highest sensitivities are achieved using ETAs due to their small cavity volume, while plasmas are not as attractive due to the high intensity of background radiation [28]. Reviews on vapor generation of individual elements are covered elsewhere [15], as are various aspects of sample preparation.

The detection system usually consists of a narrow bandpass filter or monochromator, followed by a photomultiplier tube and boxcar integrator [20]. The resulting signal is than digitized and stored for further data analysis. Recent papers highlight statistical approaches in data analysis [37], LOD determinations [38-42], and multivariate analysis [43].

Another recent spectroscopic technique is laser-induced breakdown spectroscopy (LIBS), which is sometimes called laser-induced plasma spectrometry (LIPS) [44]. Though still in its infancy, LIBS is the direct descendant of LA-LEAFS. Whereas LA-LEAFS uses two lasers to accomplish ablation and spectroscopic characterization, it was realized that the wavelength of light was not of primary importance to the ablation process. This development leads to LIBS, where a single laser pulse forms a plasma, on a solid surface or within a gas or liquid, that results in volatilization and emission from atomic species. A recent review by Russo et al. discusses the fundamentals of the ablation process in LIBS [45]. Other recent reviews discuss the instrumentation and portability aspects of LIBS [46-48], detection systems [49], and fundamental principles and spatial resolution [13,15,24,50]. Other recent reviews on LIBS analysis of aerosols and gases focus on statistical methods for data processing [51-56], matrix effects on quantitation [48,57-59], and the effects of sample inhomogeneity on measurement uncertainty [60,61].

Another review compares LIBS and other laser techniques to the “super stars” of ETA-AAS, ICP-AES, and ICP-MS [62]. Several papers make a direct comparison of LIBS to LA-ICP-MS [63-65]. In general, the advantages of LIBS is its ability to detect substrates remotely (tens of meters for solids and meters for liquids) [66], its high spatial resolution, its lack of sample preparation, its potential portability [67], and its fast analysis time. In comparison to other elemental analysis techniques, however, its LOD is quite poor [15]. It has been determined that selective volatilization plays the largest role in the variability of quantitative LIBS measurements of solid materials [68]. Though instrumentation is not the focus of this work, the authors acknowledge recent advances in hybrid instruments that utilize both LIBS and Raman spectroscopy [69-71].

Due to the relative youth of the LIBS technique, most papers are instrumentation-based, which is not the focus of this review. Most applications concern the analysis of well-characterized surfaces, such as metal alloys. Analysis of glasses and ceramics, both modern and archaeological, are popular, since the ablated mass is so minimal that the method is considered non-destructive [10-12].

Many applications of laser spectroscopic methods to the analysis of environmental solids are presented in Table 1. Of note are the recent papers evaluating the plausibility of in situ measurements using LIBS. Several papers focus on characterization of materials submerged beneath fresh or salt water. These include sediments [72], wood, and marble [73]. Ba, Mn, and Ti were detected, though calibration was hampered by shot-to-shot variability. The natural softness and roughness of the sample caused problems for laser focusing and plasma generation. Additionally, laser interrogation produced clouds of particles above the sediment surface, which scattered the laser and fluoresced light. Some detection thresholds were improved by signal processing [74]. In situ analysis of the seawater itself is discussed in the next section.

A series of recent publications from Cheam and coworkers involves measuring thallium (Tl) in sediments [94], water [95,96], and aquatic organisms [94,96,97] using LEAFS. Thallium has been scarcely studied due to its low concentrations and poor sensitivity in comparison to its more popular family members (Hg, Cd, and Pb), despite the fact that Tl possesses comparable toxicity [97]. Their investigations of coal power plants and mines suggest that coal type (rather than quantity) and/or geological contributions are responsible for the high Tl concentrations observed [95], though coal power plants and mines contain higher thallium concentrations than the other ecosystems studied [96]. Interestingly, by studying the Tl concentrations in aquatic organisms, they found that sites of maximum metal concentrations were not necessarily sites of maximum bioavailability and sediment concentrations are poor indicators of environmental distress [96,97].

A recent review by Simeonsson and collaborators discusses the use of laser spectroscopy techniques for arsenic analysis [98]. A number of atomization sources are discussed, including ICP, ETA, rhenium-coil, LA, along with hybrid techniques involving hydride generation (HG) and high performance liquid chromatography (HPLC). Limits of detection range from 4 (ICP) to 0.0003 ng/ml (HG) for aqueous samples [99-104]. Results were compared to other laser-based techniques and non-laser-based techniques, notably AFS and ICP mass spectrometry [98]. The majority of these studies, however, have been in aqueous solutions, though a few have studied diluted blood plasma [104-107] and plant tissue digests [98]. An LA-flame-LEAFS approach was applied to As on glass slides, which has yielded LODs of 0.02-0.5 μg/g ablated material [98, 102]. The uncertainty in this value is due to the uncertainty in the ablation spot diameter. Another recent review on arsenic speciation was conducted for the literature between 2000 and 2003 [108]. Others focus on the speciation of As^III, As^V and other forms of arsenic in water samples, using a variety of spectroscopic techniques [13,109-111].

LIBS was recently used to determine Si/Ca and Ca/Mg ratios, in order to distinguish between sandstone, limestone, marble, and mortar [112]. The technique was tested on the façade of a cathedral, which identified the materials of the building with good resolution over a surface of 900 m². The analysis of distant objects has been reviewed recently, focusing on experiments in which light was transmitted through the atmosphere, rather than an optical fiber [66]. The review of so-called “stand-off” LIBS discusses issues pertaining to remote measurement, including laser characteristics and optical systems associated with laser focusing and collection of fluoresced light. Applications of LIBS to non-flat [113] and moving [114] surfaces have recently been pursued. There is also a recent evaluation of common matrix effects encountered when performing LIBS analysis of geological materials [115].

Recent work has been done on the LIBS analysis of organic and biological materials [116]. An inherent shortcoming of this method is that these substances are largely composed of carbon, hydrogen, oxygen, and nitrogen. Relative signal intensities have been used to determine empirical formulas, though difficulty arises due to significant interferences from atmospheric nitrogen and oxygen. Emission from C₂ and/or CN has allowed the differentiation between organic and inorganic samples.

LIBS has also made headway in the analysis of biological samples, though only papers that are environmentally relevant are presented here. One paper studies uptake of lead into Helianthus annus [117]. Plants were hydroponically grown in different concentrations of lead acetate and dried leaves were analyzed directly using LIBS. Pb uptake was shown to alter the spatial distribution of Mn and K in leaf tissues. Another study used the upper surfaces of leaves as substrates for accumulating atmospheric aerosol [118]. Aerosols that had undergone dry deposition on Sophora japonica leaves were evaluated for their metal content using femtosecond-LIBS.

There are several recent methods for the in situ analysis of water. In two laboratory studies, LIBS was used to detect dissolved Ca, K, Li, Mn, and Na [119,120]. Pressures of up to 272 atmospheres (4000 psi) were tested, with no pressure effect on Ca or Na. The emission of Mn actually increased with increasing pressure. Additional salinity (NaCl) increased the signal for Ca and had no effect on Mn or K. A paper investigating the matrix effects in LIBS for the analysis of potassium in water droplets has a heavy bearing on seawater analysis [121]. Several other studies on water and seawater are presented in Table 2.

A promising method has recently been employed that uses wood substrates for water analysis [131]. With water absorbed into the wood, LIBS was used to quantify heavy metals such as Cr, Mn, Cu, Cd, and Pb. LOD were between 0.029 and 0.59 mg/L, which is 2–3 orders of magnitude more sensitive than LIBS without a substrate. Other recent publications utilize a substrate to provide a restricted space for the sample, such as a metal plate with 100 μm holes for the analysis of soils by LIBS [132]. The authors claim that this system reduces lost sample particles during irradiation and that atomization and excitation are enhanced due to sample confinement in the surrounding metal.

Laser-based fluorescence techniques have recently been applied to aerosol analysis. Much of this work has been spurred by tighter regulation of point sources of metals, such as furnaces and incinerators. The majority of studies concerning the bulk analysis of aerosol use X-ray fluorescence or ICP-MS, while single-particle analysis is dominated by aerosol mass spectrometry. However, the ability to perform (near) real-time analysis in the field has made LIBS and LEAFS attractive techniques. Several examples of aerosol analysis are shown in Table 3. LIBS analysis of biological aerosols (spores, pollen, etc.) suffers from the same problems as other organic and biological samples (discussed above). For this reason, research on aerosol analysis is gravitating to other techniques or LIBS in parallel with other techniques [133].

Recent advances in utilizing Raman microscopy and electron microprobe analysis (EMPA) have allowed characterization of single aerosol particles [138]. The combined techniques yield information on size, morphology, elemental and molecular composition, and molecular structure for particles as small is 500 nm in diameter. Particles were transferred from an impactor to a thin glass needle and piezo-driven nano-manipulators allowed sequential analysis of single aerosol particles by both techniques. Much of the recent literature has not focused on novel applications of LIBS to aerosols, but on technical aspects of analysis.

Laser-based spectroscopic techniques have also been applied to the analysis of trace gases in the atmosphere. Several recent reviews discuss techniques for the analysis of HO_x (OH^• and HO₂^• radicals) [139,140]. Shortly after recognizing the importance of OH^• in the atmosphere, LIF measurements of OH^• were performed in 1972 [141]. The technique has since matured significantly. LIF measurements are typically performed in reduced pressure chambers and the method is also known as fluorescence assay by gas expansion (FAGE). Specific experimental setups for LIF OH^• measurements are numerous, but all modern instruments possess a gas expansion region, a high repetition rate laser system, an electronically-gated detector (photomultiplier or microchannel plate) that is operated in photon counting mode, and a calibration system [140]. Since the lifetime of OH^• is ∼1 second, LIF offers the advantage of in situ measurement, fast analysis time, and excellent sensitivity. Selectivity is also very good. Fluorescence from other species at 308 nm, such as SO₂ and formaldehyde, are discriminated against by delayed gating. The only major interference is O₃. LIF instruments are, however, costly and have sufficient size and weight to make field measurements cumbersome (i.e., the instrument is usually housed within a small flat-bed trailer with a second trailer for the roots pump). The largest disadvantage of the technique is its need for calibration, as measurement accuracy is typically limited by the certainty of the OH^• calibration source [140]. Typical LOD for OH^• measurement during the daytime is 1.4 × 10⁵ molecules/cc for S/N of 1 and a collection time of 300 seconds [139,140]. Smaller LOD are achievable at night.

Another laser spectroscopic technique is also commonly used for OH^• measurement: long-path differential optical laser absorption spectroscopy (DOAS). Though DOAS is not covered in this review, a recent comparison between DOAS and LIF has been made [142]. LIF was found to have greater precision and time resolution, while DOAS is regarded as a primary standard due to its greater accuracy. Excellent agreement was found between the techniques, though LIF measurements were somewhat higher under high NO_x conditions. Numerous calibration instrument have been devised, including those that involve photolysis of water, measurement of hydrocarbon decay rates, and generation of OH^• through ozonolysis of alkenes [140,143,144]. Typically, several of the aforementioned methods are used. A new calibration source for HO₂^• has been devised, which uses UV photolysis of water over a TiO₂ catalyst [145]. Instrument intercomparisons have been made for ground [140], laboratory [144], and aircraft-based instruments [140]. Measurement of HO₂^• and RO₂^• radicals typically involve chemical conversion to OH^• followed by LIF detection [146,147] and new developments in this area are continuing [148].

Several recent field experiments and campaigns utilizing LIF for OH^• measurement are referenced here [149-152]. Since findings from these campaigns are typically lengthy, involve numerous measurements on a variety of compounds, and are somewhat specific to the regions they study, findings from these works will not be enumerated here.

The CRDS technique has, perhaps, found its most extensive use in atmospheric studies. The past decade has seen an exponential increase in the number of studies that use the technique. The advent of technology and the innovations of schemes that result in experimental stability, simplicity, and portability have increased its applicability for scientists and industry professionals. Its high sensitivity allows for measurement of the very weak absorptions and very dilute concentrations that are typical for trace gases. In addition, the technique's sensitivity has been used to probe the photochemistry of the atmosphere and has provided valuable detail to our understanding of atmospheric chemistry. Although there are diverse atmospheric applications of the technique, this review will focus on mainly trace gas detection, aerosol characterization, and isotope ratio measurements. Other atmosphere applications include atmospherically important chemical reaction and kinetics studies [184-186] and studies of photochemical properties of atmospherically relevant molecules [187-190]. There have been multiple reviews of the application of CRDS to atmospheric and environmental studies [158,159]. In fact this review is structured similar to that of Brown [158]; however, deliberate effort is made to focus only on the studies that have been conducted since 2005.

An exciting area of development for spectroscopic techniques based on a high-finesse optical cavity is in their biological/medical applications. Each application discussed in this subsection (CO₂ stable isotope analysis, nitric oxide and hydrocarbon measurement), is also relevant toward atmospheric trace gas studies.

Modern medicine operates on the principle that detection of different biomarkers can yield information for diagnostic purposes. One of these detection practices is through breath analysis. Breath-analysis diagnostic methods do not enjoy the same level of use as other diagnostic methods (blood test, urine test, etc.) [239]. However, the advantages for breath analysis over other techniques are multifold. Breath analysis is non-invasive and inherently safe because the subject is only required to breathe. In this section we review the application of the CRDS technique for biological sensing, mainly for human breath diagnostic methods.

Early suggestion for the potential use of CRDS for human breath diagnostics were made by Kleine et al. [240]. In their study they measured ¹³CH₄ in ambient air in real-time using cavity leak-out absorption spectroscopy (CALOS) with a tunable CW CO laser combined with a tunable microwave side-band generator. They were able to achieve detection limits in the hundred ppt ranges. A mid-infrared quantum-cascade distributed-feedback laser (DFB-QCL) used in a CRDS setup by Paldus et al. [241] showed that a detection limit of 0.25 ppb was possible for ammonia in nitrogen at standard temperature and pressure. The first application of detection using CRDS in actual exhaled breath sample was done Menzel et al. [242]. Studies on different molecules are discussed below.

The photophysical properties of closed-shell, d¹⁰, transition metal systems have gained considerable attention in recent years. Gold(I) complexes are ordinarily colorless and non-emissive [278,279]. However, they can exhibit luminescent behavior due to self association through the formation of short Au⋯Au contacts known as aurophilic interactions [280,281]. Such contacts are dominant in crystalline solids but may also occur in solutions. These aurophilic interactions are unique because Au(I) complexes involve closed shell metal centers with no apparent valence electrons to connect the Au(I) centers. Nevertheless, they still manifest the tendency to self-associate through weak interactions in the ground state that become stronger in the excited state. Relativistic effects have also been noted to influence the self-association Au⋯Au interactions, as discussed recently by Pyykkö [282,283]. The high speeds of all the electrons as they move near a heavy nucleus leads to a radial contraction of the s and p orbitals resulting in an energetic stabilization [283]. Consequently, these weak interactions are revealed via apparent changes in their PL behavior. In the frozen state, aggregate formation has been reported to increase Au⋯Au interactions, causing enhancement in the PL intensity [284,285]. The review reported by Balch in 2007 [286], covers an adequate range of topics relevant to the luminescent properties of Au(I) complexes in two-coordinate situations. Most of the Au(I) complexes have been found to luminesce as solids and in this review emphasis is placed on the solid state spectroscopy.

While the Au⋯Au interaction is important for examining the sensing potential of Au(I) complexes, the recognition of other factors is also essential for a complete understanding of their PL properties. One of these factors is the proper choice of bridging ligand. The majority of the ligands used to bridge the Au⋯Au contacts have been phosphines or P-donors [287,288], cyanides [278,289], alkynyls [290], as well as other carbon, nitrogen, and sulfur donor ligands.

In 2003, Eisenberg and Lee [291] reported gold(I) thiouracilate complexes containing bis(diphenyl-phosphino)methane (dppm) that exhibited a phenomenon known as “tribochromism”. In contrast to triboluminescence, tribochromism is a sustained change in the emission spectrum upon the initial application of pressure [291]. The Au(I) complexes reported by Eisenberg are initially weakly emitting or nonemissive. After gentle crushing, however, the compound's emission is dramatically converted to an intense blue emission at 483 nm. It was concluded that the application of pressure to the complexes induces cleavage in the weakest links of the Au(I) helix (Figure 2b). Assefa et al. [289] also observed tribochromism in the complex [(TPA)₂Au][Au(CN)₂], where TPA = (1,3,5-triaza-7-phospha-adamantane), and concluded that since powdered Au(I)-chain complexes have increased surface areas, more sites of chain termination are exposed. Therefore, surface modifications may also induce emission changes through bonding rearrangements, and through formation of localized lattice defect centers [289].

A frequently cited work in the literature is the study by Eisenberg and co-workers [292]. This investigation has served as a catalyst for most of the works conducted over the last decade employing Au(I) complexes for VOC sensing applications. The Eisenberg group reported a “switching on” of orange luminescence upon selected VOCs interactions with a dimeric Au(I) dithiocarbamate complex [Au(S₂CN(C₅H₁₁)₂)]₂. When the complex is exposed to vapors of aprotic solvents such as acetone, acetonitrile, dichloromethane, and chloroform, it exhibits structural changes where linear chains of Au atoms are formed with short intermolecular distances of 2.9617(7) Å. As a result of exposure to the VOCs, the complex becomes emissive. In the absence of these VOCs, the emission of the complex is completely quenched.

On the other hand, exposing the complex to vapors of protic solvents e.g., methanol or ethanol, provides a colorless and non-emissive crystal with a structural modification where the shortest intermolecular Au⋯Au distance is 8.135 Å. Thin films of the gold complex that were saturated in dichloromethane, exhibited an emission band at 630 nm. The absorption and emission maxima of the films exhibit little variation with the different aprotic solvents, except that when heated and dried the orange films become pale yellow and non-emissive. After the gold complex was diluted and glassed with DMF:MeOH:CH₂Cl₂, a broad emission at 563 nm was observed, which was significantly higher in energy than the thin films. Consequently, the longer chains of Au-complexes in the solid-state absorb and emit at longer wavelengths than the shorter chains in the glass since the Au⋯Au interaction is often restricted in glassed complexes.

The effect of solvent polarity on Au⋯Au interactions and the PL properties of Au(I) complexes were reported by Tzeng and co-workers [293]. The Tzeng group studied the hexanuclear complex, {[(8-QNS)₂Au(AuPPh₃)₂]}₂·(BF₄)₂ (8-QNS = quinoline-8-thiolate), which is portrayed in Figure 4. At room temperature, the solid state complex yields a single emission band at 587 nm. However, in various organic solvents intense higher-energy bands are exhibited at ∼420–480 nm, as well as lower-energy bands that show solvent-dependent features in the ∼620–680 nm spectral region. As shown in Figure 4, the polar solvents methanol, THF, and chloroform induce significant quenching of the low-energy bands in comparison to dichloromethane. In contrast, acetonitrile completely quenches the lower-energy band, while yielding the most intense higher-energy band at 477 nm. These solvent-dependent behaviors are hypothesized to arise from scrambling of the [PPh₃Au]⁺ units within the complex [293].

A similar solvent-dependency has been demonstrated by Yam and co-workers [290]. It was observed that when tetranuclear Au(I) calixcrown alkynyls are excited by 350 nm, they yield emission bands at ∼587 nm when in solution with chloroform and as frozen matrices (glass). However, in the solid state, excitation at 370 nm yields lower energy bands that are red-shifted to ∼620 nm. The respective red-shift is clearly attributed to the presence of intramolecular Au⋯Au interactions which give rise to a narrowing of the HOMO-LUMO energy gap, most likely due to orbital splittings perturbed by intraligand excited states.

Bimetallic systems involving mixed-metal interactions have also been explored recently. The “argento-aurophilic” Ag(I)–Au(I) interactions have been reported in the literature for potential VOC sensing applications [275,294-304]. For instance, recent works by Elosúa et al. have implemented Au-Ag complexes as optical fiber sensors [275,299,305]. Since both gold and silver have strong affinities for soft donor ligands, chelating phosphines have been used extensively to bridge two Au and/or Ag metal ions in close proximity [306]. Recent theoretical work by Fernández et al. describes how mixed-metal interactions enhance the dispersion forces by introducing dipolar interactions between dissimilar metals [307]. Consequently, this results in shorter metal-metal separations as compared with their respective homometallic analogues. The Fernández group [296] also characterized the structure of a [Ag(py)₃][Au(C₆F₅)₂] complex, which showed an extended ligand-unsupported chain of alternating gold and silver atoms. The short Au⋯Ag interactions are postulated to arise both from an attractive ionic(dipolar) contribution and from dispersion-type correlation effects [296]. Catalano and Horner [295] have also verified experimentally that shorter metal-metal separations exist in the mixed-metal complex [AuAg(dpim)₃]²⁺ (dpim = 2-(diphenylphosphino)-1-methylimidazole).

The Fernández group [294] also examined the properties of the extended linear chain complex {Ag₂L₂[Au(C₆F₅)₂]₂}_n (L = Et₂O, Me₂CO, THF, or CH₃CN). In this complex, pentafluorophenyl groups serve as asymmetrical bridges between the Au and Ag atoms as illustrated in Figure 5. Indeed, organic vapor inclusion into solid state complexes such as those containing Au–Ag linear chains can improve the packing within the crystal lattice by filling the voids with VOC molecules. A reversible solvent exchange often takes place when anisotropic packing adheres in a host material since vapor can be easily included into the unoccupied space in the lattice [294].

Catalano and Etogo used N-heterocyclic carbene (NHC) ligands to support closed-shell metal ion interactions [308]. Additionally, NHC ligands have been used as supports for weak Ag–Ag interactions [309-312]. In their 2005 report, Catalano and Etogo [308] observed a lack of correlation between emission energies and Au–Ag separations or the donor ability of a nitrile ligand in three nitrile-containing polymeric compounds. The origin of the emission energy shift was postulated to arise from the geometry of the metal chain and ancillary ligand electronic/steric properties rather than the metal-metal separation.

Ag(I) complexes are by far the most understudied compounds in regard to their PL behavior, despite being in the same family as Au(I) and Cu(I). A few accounts of their PL properties have been reported [313]. Omary and Rewashdeh-Omary have reports on PL (AgCN)₂⁻ complexes [314-317]. However, very few investigations have been reported that discuss the interactions between Ag(I) complexes and VOCs, and therefore remains a wide-open area for further exploration.

Fernández and co-workers have reported potential sensor applications on numerous gold-thallium complexes that depend on bimetallic interactions [307,318-325]. A polymeric {Tl[Au(C₆Cl₅)₂]}_n complex reported by Fernández et al. reacts with VOCs to yield products where the VOCs are coordinated to the Tl(I) centers [321]. The emission from this complex shifts to higher energies with increasing temperature due to increases in the metal-metal separations. The thermal expansion that results in increased metal-metal separation usually corresponds to an increase in the HOMO-LUMO gap energy [326].

The Fackler group also reported similar studies in 2003 [327], where a reversible vapochromic behavior is observed with the same {Tl[Au(C₆Cl₅)₂]}_n complex when exposed to acetone, acetonitrile, triethylamine, acetylacetone, and other VOCs. At low temperatures, the emission of this complex exhibits red-shifts when exposed to all of the listed VOCs. This shift to lower-energy is attributed to thermal contraction of the Au-Tl distances that decreases the HOMO-LUMO gap. Furthermore, Figure 6 shows the {Tl[Au(C₆Cl₅)₂]}_n crystal possessing channels running parallel to the crystallographic c-axis. The channels have diameters as large as 10.471 Å, which can easily accommodate vapor molecules into the lattice. A blue-shift occurs as the crystals are crushed to a powder which is shown in Figure 7. This phenomenon is attributed to the reduced spatial confinement of the electrons, which increases the oscillator strength between orbitals, thus demonstrating the quantum size effect.

Fernández and co-workers are also one of the first groups to synthesize mixed Au(I)-Cu(I) complexes. The photophysical properties of a [Cu{Au(C₆F₅)₂}(N≡CCH₃)(μ₂-C₄H₄N₂)]_n (C₄H₄N₂ = pyrimidine) complex were demonstrated for VOC sensing [302,328]. Since Cu(I) has a large affinity for nitrogen-donor ligands, more ligand variations and experimental studies are possible for mixed-metal systems.

Copper(I) compounds, which previously did not receive much attention [329,330], are gaining interest for VOC sensing as well as being considered for use in host-guest systems [331]. Ford and co-workers [266,332] have reported earlier investigations of PL Cu(I) complexes, where vapoluminescence behavior was observed in two solid [CuI(4-pic)]_x complexes, one existing in a tetrameric form (x = 4) and the other in a polymeric form (x = ∞). Upon exposure to toluene vapor, the polymeric form exhibits a vapoluminescent shift in emission from 437 to 580 nm (Figure 8). In fact, toluene exposure results in a complex transformation from the polymeric form to the tetrameric form. However, the process is reversed when exposed to n-pentane vapors for three hours. Although the low cost of Cu(I) compounds makes them advantageous in practical sensing applications, the slow response times and dependability on particle dimensions compromises their suitability. For instance, the initial complex transformation is faster when it is finely powdered. For further detailed studies on Cu(I) complexes, the Ford group completed a thorough review outlining their PL properties [333].

More recently, Omary and co-workers reported on the photophysical properties of a trinuclear Cu(I) complex {[3,5-(CF₃)₂Pz]Cu}₃, referred to as Cu₃ [334]. The emission behavior of the complex is affected by several factors including temperature, and exposure to benzene, toluene, and acetonitrile solvents. In addition, the Omary group introduced the concept of “concentration luminochromism”, a phenomenon that takes place when multiple visible emissions can be tuned by controlling the concentration of the Cu₃ complex.

Platinum (Pt) complexes have been one of the most studied transition-metal complexes since the 19^th century and even well before the advent of coordination chemistry [335]. Ultimately, the ability of Pt(II) systems to selectively interact with specific compounds via vapoluminescence has served as a motivation for several studies [273,336-346]. Che and co-workers prepared four trinuclear [(CˆNˆN)₃Pt₃(μ₃-L)]³⁺ complexes by treating them with phosphine ligands [347]. They reported that the absorption and emission energies of oligmeric cyclometalated Pt(II) complexes can be tuned to a large extent by the proper choice of tethering phosphine ligands. These Pt complexes are shown to serve as amplifiers for subtle concentrations of VOCs.

In contrast to being subtle, Pt(II) complexes exhibit some of the largest vapochromic shifts ever seen among transition metal complexes. Du and co-workers [338] reported on a [Pt(TPPPB)Cl]Cl (TPPPB = 1-terpyridyl-2,3,4,5,6-pentaphenylbenzene) complex as showing high selectivity for VOCs such as methylene chloride, ethanol, ethyl acetate, and acetonitrile. Upon exposure to methylene chloride vapors, the emission of the [Pt(TPPPB)Cl]Cl complex exhibits blue-shifts from 654 to 514 nm. The resulting solid state PL spectra are shown in Figure 9. In the figure, the [Pt(TPPPB)Cl]Cl complex is referred to as 5-R in the absence of VOC exposure, and 5-G in the presence of the VOC vapor. The 5-R complex possesses Pt⋯Pt distances of 3.30 and 3.34 Å, whereas the 5-G complex contains significant Pt⋯Pt interactions with distances of 3.9092(9) and 4.5483(11) Å.

In 2002, Kato et al. reported on a dinuclear platinum(II) complex that exhibited vapochromic changes in the presence of acetonitrile and ethanol [348]. Two geometrical isomers containing [Pt₂(2,2′-bypyridine)(pyridine-2-thiolate)] ions were synthesized, one being a syn isomer in the form of dark-red crystals, and the other being an anti isomer in the form of orange needlelike crystals. The Pt⋯Pt separations for the syn and anti isomers are 2.923(1) and 2.997(1) Å, respectively. The intriguing observation was that the syn isomer crystals, which are initially dark-red, when left standing in air at room temperature become light red and exhibit red luminescence at 644 nm with a lifetime of 170 ns. Figure 10 shows images of the crystals after exposure to acetonitrile or ethanol vapor in which they become dark-red again with a shift in emission towards the infrared region at 766 nm.

The vapor-induced luminescence switching is evident in the syn isomer crystals; however, the sensitivity is reduced with increasing size of the organic vapor molecule. For instance, the syn isomer is less sensitive to isopropanol and exhibits no luminescence at all when exposed to tert-butanol despite both of them having similar vapor pressures to ethanol [348]. The source of vapochromism is attributed to the crystal structure of the syn isomer, where a co-planar-arrangement reveals a channel within its lattice that can hold one acetonitrile molecule per complex. In fact, X-ray analysis indicates several peaks of electron density within the channel, confirming that there are no distinct interactions between the main body of the complex and the acetonitrile molecule. The channel simply provides a pathway for organic vapors to easily penetrate the crystal. As a result, vapoluminescence occurs as Pt⋯Pt interactions are induced by the channels filling up with specific organic vapor molecules. The orange anti isomer crystal does not contain channels within its lattice, hence providing evidence for its lack of vapoluminescence.

Since the early discovery of the Magnus' salt, [Pt(NH₃)₄][PtCl₄], in 1828, optical investigations of Pt(II) salts continue to attract attention [267, 336, 343]. Connick and co-workers reported the vapoluminescent behavior of Pt(II) salts [349]. Cl⁻ and PF₆⁻ anions of a Pt(Me₂bzimpy)Cl⁺ (Me₂bzimpy = 2,6-bis(N-methylbenzimidazol-2-yl)pyridine) complex were studied. PL studies conducted on the Cl⁻ salt in glassy solutions reveal emission bands at 545, 590, and 680 nm whose intensity depends on the complex concentration. The 680 nm band gains considerable intensity with increasing complex concentration, indicating possible formation of emissive aggregates. At room temperature, the Cl⁻ salt in the solid state exhibits a broad band near 670 nm that blue-shifts and becomes sharper upon exposure to methanol vapor (Figure 11b).

When the Cl⁻ salt is cooled, the emission bands sharpen further and red-shift to 685 nm both before and after exposure to methanol, indicating that thermal effects cause lattice contractions which shorten the Pt⋯Pt interactions. The solid-state emission studies of the Cl⁻ salt suggest that sorption of methanol vapor increases the Pt⋯Pt interactions, which possibly changes the orbital character of the lowest emissive state. Similar behavior is observed for the PF₆⁻ salt when exposed to acetonitrile vapor except emission maxima are more red-shifted due to stronger Pt⋯Pt interactions.

More recently, Connick and co-workers have also demonstrated how vapochromic salts of Pt(II) systems may present advantages over other vapochromic Pt(II) systems in terms of selectivity, color change, and response speed. Since these characteristics are drastically dependent on the counteranion, this provides a strategy for tuning the response of the complex [350]. At the molecular level, anions and VOC molecules fill the voids between the columns of cations in vapochromic Pt(II) salts. Similarly, VOC molecules line channels along the c-axis, also providing a possible route for diffusion of vapors in and out of the host lattice.

Pt(II) coordination compounds are also among the most prevalent form of extended linear chain (ELC) compounds used in VOC sensing [351]. Coordination complexes with Pt(II) centers are often substituted by palladium centers due to their similar chemical properties. However, the two atoms show considerable differences in the manner of their ELC structures [352] and crystal packing [353]. As a result, the capacity for Pt and Pd complexes to accommodate VOCs through lattice channels can vary. The lack of reversibility is a major impediment in some of the Pt(II) ELC complexes that may have the potential for VOC sensing. Drew and co-workers [351] investigated a Pt^II(CN-i-C₃H₇)₂(CN)₂ complex that shows a considerable spectroscopic change upon exposure to benzene. The selectivity of this crystalline complex for benzene exhibits remarkable shifts from a yellow emission at 558 nm to a blue emission at 484 nm upon the formation of a benzene solvate. However, the removal of the benzene from the resulting solvate yields a 50% reduction in both the absorptivity and the emission intensity. Furthermore, a repeated exposure of the same sample to benzene does not match the initial emission spectrum of the benzene solvate. The loss of reversibility is attributed to the molecular structure of the Pt^II(CN-i-C₃H₇)₂(CN)₂ complex because it lacks the necessary molecular channels to accommodate the penetration of benzene into the lattice. Consequently, the exposure of benzene forces individual crystallites to open, destroying the long-range order of the lattice. This renders crystalline degradation that can only be reversed by exposure to an acetonitrile precursor. Indeed, previous studies have shown that the choice of precursor molecules can also effect Pt⋯Pt interactions since they can vary the size and structures of nanoscale aggregates [344].

Other metal complexes such as tin(II) salts exhibit remarkable PL properties that are often quenched upon exposure to organic vapors. Baldauff and Buriak reported an optical sensing array comprised of four luminescent Sn(II) salts that are selectively quenched by amine vapor [354], where the quenching events vary in terms of their reversibility. The study consisted of the sulfate, methanesulfonate, triflate, and fluorophosphate salts of Sn(II), which are shown in Figure 12. The quenched blue emission from the fluorophosphate salt is consistently reproducible after simple N₂ purging. In contrast, the sulfate and methanesulfonate salts are only partially regenerated after exposure to trifluoroacetic acid (TFA) vapors, while the triflate salt yields no reversibility after N₂ purging or TFA exposure. Hence, Baldauff and Buriak have opened the door for more investigations to utilize Sn(II) compounds for optical chemical sensing.

Navale and Mulla [355] reported on the use of Sn as a dopant in ZnO crystals for vapor sensing. The inclusion of Sn into the ZnO crystals induces a significant morphological change that causes a reduction of sensitivity towards acetone vapors. Also as a result of the doping, the emission band shifts dramatically from the UV region to a green emission. Gu and co-workers have also reported potential sensing applications of luminescent SnO₂ compounds in the form of thin films [356] and nanoparticles [357,358].

Ruthenium(II) complexes have also been considered for sensing polar organic solvents [359]. McGee et al. has reported a ruthenium [Ru(5,6-Me₂Phen)₃]tfpb₂ (5,6-Me₂Phen) = 5,6-dimethyl-1,10-phenanthroline; tfpb⁻ = tetrakis(bis-3,5-trifluoromethylphenylborate) complex that was examined for the sensing of benzene and oxygen [265]. The solid Ru(II) complex demonstrates a very rapid and reversible vapochromic shift of emission from 572 to 558 nm when exposed to benzene alone, while exposure to oxygen alone results in PL quenching. In contrast, when exposed to benzene and oxygen simultaneously, the Ru(II) crystals uptake benzene molecules preferentially and the diffusion of oxygen into the lattice is restricted. Therefore, the emission intensity does not exhibit significant quenching. Nevertheless, upon the removal of benzene from the system, the oxygen-induced quenching is restored. The efforts of the McGee group are ongoing in an attempt to meld these two sensor mechanisms into a single functioning system. Coincidentally, some luminescent Ru(II) complexes have also been reported as being highly sensitive in the presence of moisture [360].

Luminescent iridium complexes have also been studied for their unique selectivity of VOCs. Liu et al. [361] reported a vapoluminescent cyclometalated heteroleptic iridium complex [iridium(III)-bis(2-phenylpyridinato-N,C2)(quinoxaline-2-carboxylate); PIr(qnx)] that was synthesized into two forms (black and red forms). The black form is weakly luminescent at 692 nm, with a short lifetime of 43 ns. The black form also transforms into the red form when exposed to acetonitrile or propiononitrile vapor, but exhibits no response when exposed to 15 other VOCs. The red form displays intense PL at 654 nm with a lifetime of 130 ns, demonstrating selectivity for acetonitrile vapor. Photographic images and solid state PL spectra are shown in Figure 13. The acetonitrile selectivity of the black form is attributed to the weak interaction of the left-over solvent molecules within the lattice. Since the black crystals were grown from ethanol/chloroform, the oxygen atoms of the trapped ethanol within the lattice are held at a distance of 3.27 Å from the hydrogen atoms of the PIr(qnx) complex, allowing ethanol molecules to be easily substituted by acetonitrile.

Reports on the use of cobalt for VOC sensing applications are limited. Beauvais et al. [362] reported on clusters of [Re₆Q₈(CN)₆]^4- (Q = S, Se) which were used to space out hydrated Co(II) ions creating porous materials that display dramatic color changes upon exposure to certain organic solvents. Complexes of [Re₆Q₈(CN)₆]^4- anion with various Co(II) ions such as [Co₂(H₂O)₄]⁴⁺, have flexible structural frameworks that can expand to accommodate the incoming solvent molecules.

Similar vapochromic shifts in absorption spectra were also observed for some nickel(II) complexes. Baho and Zagarian [363] have reported a series of Ni(II) complexes coordinated with a diphenyl-(dipyrazolylmethane) (dpdpm) ligand. Upon exposure to acetonitrile vapor, the solid state complexes exhibit vapochromic shifts from green to light blue in approximately eight minutes. The shift is attributed to a reversible coordination of acetonitrile at the Ni center. Thermochromic behavior is also observed in these Ni(II) complexes upon exposure to higher temperatures [363].