Ultrathin Calcinated Films on a Gold Surface for Highly Effective Laser Desorption/Ionization of Biomolecules

Abstract

We report a nanoscale calcinated silicate film fabricated on a gold substrate for highly effective, matrix-free laser desorption ionization mass spectrometry (LDI-MS) analysis of biomolecules. The calcinated film is prepared by a layer-by-layer (LbL) deposition/calcination process wherein the thickness of the silicate layer and its surface properties are precisely controlled. The film exhibits outstanding efficiency in LDI-MS with extremely low background noise in the low-mass region, allowing for effective analysis of low mass samples and detection of large biomolecules including amino acids, peptides and proteins. Additional advantages for the calcinated film include ease of preparation and modification, high reproducibility, low cost and excellent reusability. Experimental parameters that influence LDI on calcinated films have been systemically investigated. Presence of citric acid in the sample significantly enhances LDI performance by facilitating protonation of the analyte and reducing fragmentation. The wetting property and surface roughness appear to be important factors that manipulate LDI performance of the analytes. This new substrate presents a marked advance in the development of matrix-free mass spectrometric methods and is uniquely suited for analysis of biomolecules over a broad mass range with high sensitivity. It may open new avenues for developing novel technology platforms upon integration with existing methods in microfluidics and optics.

The development of soft ionization techniques such as matrix assisted laser desorption ionization (MALDI) has greatly changed the field of mass spectrometric analysis¹. MALDI-MS benefits from its salt tolerance, simplicity of mass spectra, and broad mass range. Because of the interference from matrix-related ions in low mass range, MALDI-MS is seldom applied to the analysis of low-molecular weight compounds (below 600 Da). This limitation has hampered its wide application in important research fields such as drug discovery and biotechnology, where small molecule detection and identification is of high significance². In recent years, various methods leading to direct desorption/ionization without organic matrices have been extensively explored, and surface-assisted laser desorption/ionization (SALDI) has gained considerable attention and has found a broad range of applications in environmental, genomics and proteomics fields owing to its attractive features of simple sample preparation and low background ions^3–5. SALDI relies on direct absorption of UV laser light by the substrate or its coating that lead to molecular desorption and subsequent ionization. A range of materials have been investigated for their effective use in SALDI. Nanomaterials in different forms including inorganic powder^6–8, nanowire⁹, nanotubes¹⁰ and porous thin structures^11–14 have been tested as alternatives to organic matrices. Metallic nanoparticles are another family of materials that have been heavily explored^15–17. These materials show promising results but also exhibit limitations and have suffered from problems such as inhomogeneous deposition, molecular degradation and interference from metal cluster ions^17–19. Among all substrates studied so far, porous silicon (also referred to as desorption/ionization on silicon, DIOS)^{20, 21} is the most significant and well-utilized as it offers highly effective ionization due to strong UV adsorption and heat transfer. DIOS substrates are typically prepared by an electrochemical etching procedure. Yet DIOS-MS has limited upper mass range, and the surface is susceptible to oxidation deactivation and requires stringent control of surface physical properties, which must be realized through careful selection of the silicon type and etching conditions,^{22, 23} whereas nonuniform surface structures can substantially deteriorate the performance^{20, 23}. Recently a new surface-based ionization method, nanostructure-initiator mass spectrometry (NIMS)²⁴, has been reported that uses hydrophobic liquid compounds (initiators) to facilitate and improve the desorption of analytes on the DIOS substrate.

We describe here a new class of materials based on laser induced electron-phonon interactions for effective desorption/ionization and thus matrix-free mass spectrometric analysis of a range of biomolecules. Our approach utilizes a nanoscale, glass-like silicate film fabricated on a thin gold substrate through a layer-by-layer (LbL) deposition/calcination process (Figure 1). Silica (SiO₂) and silicate-based materials have been shown to have potentials for SALDI-MS analysis of small molecules^{12, 25}. These methods utilize either a sprayed coating of a sol gel solution or a homogenized particle suspension to load the silica to the sample stage for enrichment of samples in SALDI analysis. The thickness of these layers is in the range of 500–1000 μm and the uniformity is difficult to control.¹² The LbL/calcination process, on the other hand, generates a vastly different substrate and allows for precise control of the coating thickness and porosity in the nanometer scale, which is crucial to this ionization method. In this work, the effectiveness of calcinated films on gold for SALDI-MS detection of amino acids, peptides and proteins has been systematically investigated and the substrates are characterized by a number of techniques including scanning electron microscopy (SEM), atomic force microscopy (AFM) and contact angle measurement to understand the relationship between surface property and performance. The ultrathin glassified coating is stable and has high durability, and has a number of other advantages including low cost, well-defined surface property, reusability and ease of preparation and functionalization. In addition, the photonic properties of the thin gold substrate may allow for multiple modes of detection at the same surface and development of new hyphenated technologies.

Figure 1.

A schematic of SALDI-MS detection with the calcinated silicate film and its fabrication by LbL deposition of PAH and sodium silicate (water glass) on a gold surface.

EXPERIMENTAL SECTIONS

Materials and instrument.

3-Mercaptopropionic acid (3-MPA), poly(allylamine hydrochloride) (PAH), α-cyano-3-hydroxy-cinnamic acid (CHCA), [Sar¹, Thr⁸]-angiotensin II (MW = 956.1), neurotensin (MW = 1672), insulin b chain (oxidized, MW=3495.9) and cytochrome c from bovine heart were purchased from Sigma-Aldrich (St. Louis, MO). Sodium silicate (SiO_x), citric acid, trifluoroacetic acid (TFA), L(+)-lysine monohydrochloride, L(+)-arginine, L-histidine and acetonitrile were from Thermo-Fisher Scientific (Pittsburgh, PA). Stainless steel tape (SST) was purchased from https://www.labelvalue.com/ (Tampa, FL). Water was purified by a Milli-Q system. All other reagents were analytical grade and used without further purification.

Preparation of thin Au layer on substrates.

Au surface was fabricated by e-beam deposition of a 46-nm thick gold layer onto pre-cleaned SST and glass slides. 2-nm Cr film was pre-deposited on glass as an adhesion layer before Au deposition to enhance stability of the Au film on the substrate.

Preparation of nanoscale calcinated films.

Preparation of calcinated silicate layers on gold substrates was carried out using a previously published procedure by our group.^{26, 27} In brief, cleaned gold substrates were immersed in a 5 mM 3-MPA ethanol solution overnight, followed by extensive rinsing with ethanol and DI water. PAH (1 mg/mL, pH 8.0) and sodium silicate solution (22 mg/mL, pH 9.5) were alternately deposited to the surface by spray bottles with DI water rinsing between each spray. This process was repeated until the designated number of layers was reached while surface plasmon resonance (SPR) was used for monitoring of polyelectrolyte deposition and control of film thickness. Finally, deposited substrates were calcinated in a furnace by heating to 450 °C at a rate of 17 °C per min and brought to room temperature after 4 h.

Sample preparation for MS analysis.

The stock solution for peptides was prepared by dissolving [Sar¹, Thr⁸]-angiotensin II and neurotensin in 50% acetonitrile (ACN) to a concentration of 200 μM, respectively. CHCA solution (10 mg/mL) was prepared in 60% ACN/water solution containing 0.1% TFA. When CHCA was used as the matrix, the sample solution was prepared in a 1:10 ratio of peptide solution to CHCA. For MALDI-MS analysis, 1.0 μL of sample solution was deposited onto the MALDI sample plate and dried in vacuum prior to MS detection.

Calcinated substrates were first washed by DI water and ethanol, and dried by compressed air. The cleaned substrates were attached onto an MALDI plate by adhesive polyimide tape before sample deposition. Amino acids, peptides and proteins were dissolved in a 60% ACN/water solution containing 0.1% TFA and 10 mM citric acid. Aliquots (0.5–1.0 μL) of sample solution were deposited onto the calcinated surface and allowed to dry in air before SALDI-MS analysis. For reusability test, the calcinated surface was regenerated by washing with 0.5 M ammonia solution, D.I. water and pure ethanol alternately for 3 times.

Scanning electron microscopy (SEM) and atomic force microscopy (AFM).

Scanning electron microscopy (SEM) images were obtained by a Philips XL30 FEG scanning electron microscope system. The SEM measurements were carried out with a beam power of either 5 or 20 kV with magnification ranging from 10× to 80000×. AFM images were collected by a Veeco Dimension 5000 atomic force microscope (Santa Barbara, CA) with manufacturer-provided software. All images were obtained in the tapping mode, and RMS surface roughness values were obtained by averaging multiple 5 μm² areas across the entire substrate at a scan rate of 1.5 Hz.

Contact angle measurements.

Contact angle measurements were performed on a home built device with deionized water (1 μL). The images for water droplets on substrate were collected by a computer controlled 12-bit cooled CCD camera. All measurements were made in ambient atmosphere at room temperature.

LDI-TOF MS.

Laser desorption and matrix assisted laser desorption/ionization mass spectra were obtained by using Voyager-DE STR MALDI-TOF mass spectrometer (Applied Biosystems, USA) operating in positive reflectron mode. The mass spectrometer is equipped with a pulsed nitrogen laser operated at 337 nm with 3 ns-duration pulses. The accelerating voltage, grid voltage and extraction delay time were set as 20 kV, 65% and 190 ns, respectively. MS spectra were acquired as an average of 100 laser shots.

RESULTS AND DISSCUSION

Fabrication of Nanoscale Gold/Calcinated (Silicate) Chips.

Figure 1 shows a schematic of SALDI-MS detection with the calcinated chip and its fabrication with LbL deposition of poly(allylamine hydrochloride) (PAH) and sodium silicate (water glass) on an Au-covered stainless steel tape (SST). The fabrication was based on our previous work to build a nanoscale silicate glass layer on Au for SPR measurements.^{26, 27} The layer thickness and other properties of the thin film were monitored with SPR, which was performed on a glass slide based Au substrate under identical deposition conditions. A linear relationship was found between SPR angle shift and deposition layer number, demonstrating a uniform growth of PAH/silicate layers (supplement). The thickness of PAH/silicate layers can be calculated by fitting to theoretical reflectivity curves from the Fresnel equations with an average refractive index (RI) of 1.455 for one PAH/silicate layer²⁷. After achieving designated thickness and calcination at 450 °C, organic components are removed from the multilayer structure, leading to a reduction of thickness in the film, as indicated in the SPR sensorgram. Using this method, the thickness of the glassified layer can be controlled at 1 nm resolution. From the SPR angular shift, the 15-layer film after calcination has an average thickness of ~20 nm.

SALDI-MS on Nanoscale Calcinated Surface.

The SALDI analysis on nanoscale calcinated surface was carried out with two peptides, [Sar1, Thr8]-angiotensin II (MW 956.1) and neurotensin (MW 1672). Figure 2 a and b show mass spectra of the peptides on the calcinated surface and the MALDI analysis with CHCA matrix for comparison. SALDI on the calcinated surface demonstrated a very clean mass spectrum free of major noise peaks. Peptide protonated ions were dominant in the spectrum with little or no fragmentation, which clearly demonstrated excellent performance of this surface in MS analysis. For MALDI-MS with CHCA, a high background noise appeared in the low-mass region. Matrix related ions, such as [CHCA+H]⁺ at m/z 190, [CHCA+Na]⁺ at m/z 212, [CHCA+K]⁺ at m/z 228 and [2CHCA+H]⁺ at m/z 379, dominated the spectrum. By contrast, SALDI showed only a few peaks in the low-mass range, generated from ions of citric acid adducts (□) and impurity (●). The detection limit for the two peptides was at the fmol level (100 fmol [Sar1, Thr8]-angiotensin II gave S/N ~38), pointing to a sensitive detection of peptides with the chip.

Figure 2.

Mass spectra for peptides and amino acids. (a) A peptide mixture on calcinated glass surface; (b) A peptide mixture with CHCA matrix; (c) An amino acid mixture on calcinated glass surface; (d) An amino acid mixture with CHCA matrix. Peptide mixture: [Sar¹, Thr⁸]-angiotensin II MW = 956.1) and neurotensin (MW = 1672), 20 pmol each with 10 mM citric acid; Amino acid mixture: Lys, His and Arg, 60 pmol each with 10 mM citric acid. ●: impurity ions; □: citrate related ions; *: impurities in amino acid samples; ■: CHCA fragment ions. NL: normalized level.

The low background in the mass window below 500 Da suggests this calcinated surface can be useful for analysis of small molecules. Figure 2 c and d show mass spectra for three amino acids of Lys, His and Arg on the calcinated surface and with CHCA matrix, respectively. In the presence of CHCA, matrix adducted ions dominated the spectrum while sample ions were swamped by CHCA fragment ions including [CHCA-H₂O+H]⁺ at m/z 172 and [CHCA-CO₂ +H]⁺ at m/z 146, which are very close to ArgH⁺ (m/z 175) and LysH⁺ (m/z 147). On the calcinated surface, sample ions were dominant in the spectrum with much improved noise level. The ion intensity was high and signals related to citric acid adducts and impurities appeared with low abundance. The signal-to-noise ratios were improved by about 1.7 and 1.3 times for ArgH⁺ and LysH⁺, respectively, as compared to those in MALDI-MS. The results clearly demonstrate the effectiveness of the calcinated surface for MS analysis of small molecules. This feature would be particularly useful in metabonomics study and drug discovery where high throughput screening of small compounds is constantly required²⁸.

The feasibility of using the calcinated surface for SALDI-MS analysis of large molecules was also explored. Figure 3 shows the SALDI mass spectra of insulin b chain (oxidized, MW 3495.9) and cytochrome c (from bovine heart, MW 12,327). A very clean background was achieved for SALDI analysis of these large biomolecules, especially for insulin b chain. In comparison, matrix related ions were found in the mass region higher than 900 Da when CHCA was used (inset). In addition, several multi-sodium adducted ions for the peptide, including [M+Na]⁺, [M-H+2Na]⁺, [M-2H+3Na]⁺, [M+−3H+4Na]⁺ and [M-4H++5Na]⁺ were observed with CHCA matrix due to existence of two sulfonic groups in the peptide framework. SALDI with calcinated surface, on the other hand, produced only protonated sample ions and single alkali metal adducted ions, including [M+Na]⁺ and [M+K]⁺. The ion intensity ratio of protonated ions to single sodium adducted ions was about 4.7, which was much higher than 2.8 in the MALDI-MS with CHCA. This result suggests that SALDI on the calcinated silicate surface, with assistance of citric acid, tends to produce protonated sample ions and can effectively suppress the generation of alkali metal adducted ions,¹⁴ and therefore greatly simplify the mass spectrum for peptide identification. Cytochrome c can also be identified by using the calcinated surface with SALDI. The protonated pseudomolecular ions ([M+H]⁺) and the double charged ions ([M+2H]²⁺) of the protein were found, showing that the effective mass range of the substrate spans a large mass window that includes both amino acids and small proteins.

Figure 3.

Mass spectra for peptide and protein with SALDI on the calcinated glass surface. (a) Insulin chain b, 20 pmol with 10 mM citric acid; the inset spectrum was obtained with CHCA matrix on steel MALDI plate under the same conditions; *: CHCA related ions; (b) cytochrome c from bovine heart, 40 pmol in 10 mM citric acid.

In addition to large effective mass range, calcinated substrates are highly stable and have exhibited excellent long-term durability. There was no detectable loss of material in SPR spectroscopic study of incubation of the chips for hours with different buffers including Tris-HCl, NaCl and PBS. SALDI activity and performance of the substrates have no significant change after storage in air for months. Common organic solvents, such as methanol, ethanol, acetonitrile and dichloromethane, have been tested for sample deposition and surface cleaning without showing any negative effects. Additionally, the surface can be repeatedly used as many as 10 cycles with minimal loss in ionization efficiency, and readily modified by silanes without deterioration on SALDI-performance (supplement). These attractive features provide variety of choices to further facilitate LDI activity, sample deposition, and selective capture of analyte by tailoring surface properties with chemical modification²⁹.

Understanding performance-determining factors on calcinated surface.

Experimental parameters that affect LDI on calcinated surface were investigated to understand the process and improve the performance. In the case of DIOS-MS, it has been proposed that properties of porous silicon such as UV absorption, surface morphology and thermoconductivity play important roles in the LDI process.^{22, 30–32} However, silicate has no strong absorption at 337 nm of the N₂ laser and it is clear that LDI on this surface was not a direct result of UV-absorption of the glassy layer since silicate film fabricated on glass slides was SALDI inactive and yielded no signal in MS detection. In addition, we found neither bare glass cover slips nor HF-etched glass cover slips showed any activity for SALDI. These results clearly indicate that the Au layer is essential to the LDI process. Interestingly, bare gold surface was not effective for inducing LDI. Figure 4 shows mass spectra generated from SALDI on Au-covered SST and calcinated Au-covered SST chips. LDI efficiency remained very low for Au-covered SST. Ion intensities of protonated ions for [Sar¹, Thr⁸]-angiotensin II (M1) was only about 563 counts (Figure 4a). In comparison, the calcinated substrate showed a much enhanced LDI for [M1+H]⁺ with ion intensity increasing to 2170, which is an almost 4-fold increase (Figure 4b). SALDI-MS for neurotensin (M2) gave similar results. Furthermore, abundant alkali-adducted analyte ions, including [M+Na]⁺ and [M+K]⁺, were produced on the calcinated surface and showed higher ion intensities relative to the protonated ones. The alkali metal ions are thought to originate from sodium silicate in the calcinated film, suggesting some degree of ion exchange and charge separation are involved in the LDI process.³¹

Figure 4.

SALDI-MS analysis of two peptides on different substrates. (a) Au-covered SST; (b) calcinated film on Au-covered SST; (c) calcinated film on Au-covered SST with 10 mM citric acid. Conditions: samples: [Sar¹, Thr⁸]-angiotensin II (MW = 956.1) and neurotensin (MW = 1672), 20 pmol each. Spectrum (c) contains 10 mM citric acid in the sample. (*) Fragment ions generated from analytes. NL: normalized level.

We noted fragmentation of the sample ions on the calcinated chip appeared to be excessive. To suppress the fragmentation and increase production of protonated ions, an external proton donor was utilized. Figure 4c shows the performance of SALDI on calcinated films with addition of 10 mM citric acid in the peptide solution. A highly clean background in the mass spectrum was obtained, and production of alkali-adducted ions and fragment ions was highly suppressed. Protonated analyte ions dominated in the spectrum, and ion intensities for [M1+H]⁺ and [M2+H]⁺ increased by 4.4 and 15.8 times with citric acid, corresponding to nearly 8 and 30 times increase of signal-to-noise ratio, respectively. No obvious improvement of LDI was observed though on the bare Au-SST surface with the external proton donor. The concentration of citric acid in sample solution was optimized for LDI and ion abundance of protonated analyte ions was found to increase with the increase of citric acid under 10 mM concentration (supplement). It became difficult though to obtain an MS signal from sample spots when the concentration of citric acid exceeded 20 mM. When 100 mM citric acid was used, the MS response completely disappeared. This may be attributed to co-crystallization of citric acid with analytes on the calcinated surface, which forms a thick layer on top of the substrate and thus results in reduction of the LDI efficiency. The increase of sample-spot thickness may also cause problem in thermal desorption of the deposited components due to limited depth of laser penetration through the film.³¹

The effect of laser fluence on LDI on calcinated chips was investigated. The laser threshold on this surface was determined to be 1920 (a.u.), which was about 320 units higher than that for MALDI with CHCA matrix. In comparison, much higher laser fluence is required to achieve MS signals for peptide in DIOS and SALDI with porous alumina relative to CHCA matrix.³³ The lower laser fluence as compared to other substrates demonstrates a higher LDI efficiency on the calcinated surface. We further found that ion signal increased with laser intensity, reaching the climax at the laser intensity of 2250. The signal then decreased rapidly with the application of higher laser fluence (supplement). The decline of ion signal may be attributed to the damage of nanoscale calcinated layer by rapid heating at high laser fluence. This phenomenon was also observed in DIOS³⁰ and SALDI on metals.¹¹

Surface characterization and possible mechanism.

Surface roughness has been suggested to affect LDI on solid surfaces.^{33, 34} Figure 5 shows the SEM images of the calcinated surface covered with 8, 15 and 20 layers of PAH/silicate. A relative smooth surface was observed on the bare Au except repeated ridges and wrinkles arising from blunt irregularities of the SST. These irregular structures were also found on the calcinated film-covered SST substrates. For surfaces covered by the calcinated film, a relative rough superficial layer was obtained and no obvious fractures were found. Importantly, a porous structure in the nanometer scale (pore size less than 20 nm) was observed with high magnification (Figure 5b inset). These pores may come from removal of organic components by calcination and localized shrinkage of the silicate layer after annealing. As more layers of silicate were deposited, the surface roughness increased and became dominating for 15 and 20 layers films with more nanometer-sized islands formed.

Figure 5.

SEM images of calcinated surface with different layers of silicate on gold surface deposited on an SST tape; (a) bare Au surface, (b) 8 layers of silicate, (c) 15 layers of silicate, (d) 20 layers of silicate. The scale of the bar in the images is 1 μm, expect for image on the right top of (b), which is 100 nm.

AFM was also used to examine the substrates (Figure 6). The bright streaks, which appeared across the images, were the ridges and wrinkles observed in SEM images (Figure 5). The surface roughness (RMS) for the calcinated film on an SST substrate was much higher than that for the same film on glass slides. For instance, the RMS for surface with 5 layers of silicate on SST was 21.3 ± 0.6 nm, while the RMS value for 6-layer silicate on glass slide was less than 4 nm. Additionally, increased RMS values were observed on substrates with more calcinated layers, where the RMS values for 15-layer and 20-layer silicate were 27.7 ± 0.6 nm and 44.3 ± 2.3 nm, respectively. The increased RMS appears to stem from the formation of nanostructured islands, which were observed with SEM.

Figure 6.

AFM images of calcinated surface with different layers of silicate on gold substrate deposited on an SST tape; (a) 5 layers of silicate, (b) 15 layers of silicate, and (c) 20 layers of silicate.

Surface hydrophobic property of SALDI-substrates plays important role in desorption/ionization of analytes^22,35. Contact angle measurements were carried out to evaluate the surface hydrophilicity of the calcinated film with different numbers of deposited layers (Figure 7). Bare Au-SST surface showed a contact angle of 68°. Sodium silicate glass is known for its high hydrophilicity and we expected that surface would become more hydrophilic with increasing number of silicate layers. However, our results show that the contact angle value increased slightly from 68° to 73° with the first 1–5 layers of (PAH/silicate)_n deposition. This initial moderate increase of hydrophobicity might arise from introduction of free silanol groups at the surface.³⁶ When the first several layers of silicate were coated, a single-layer structure film with random uncovered areas was formed due to uneven growth of polyelectrolytes. Silanol groups exposed at the surface exist as free hydroxyl groups, which lowers the affinity of water to the surface. As more layers of silicate were coated, the surface hydrophilicity increased remarkably. When 20 layers of silicate were fabricated, the contact angle was only 17° (Figure 7). The silanol groups in multilayer structure are likely converted to hydrogen bonded hydroxyl groups,³⁶ which promote water physisorption and therefore increase the hydrophilicity. In addition, the increase in surface roughness and sodium content for multilayer silicate may also contribute to the increase of hydrophilicity of the calcinated substrate. LDI on 15-layer and 20-layer silicate showed 5.8 and 6.3 times of enhancement in terms of ion abundance for [Sar1, Thr8]-angiotensin II relative to that on 5 layers of silicate (supplement Figure S7). The results seem to verify that rough surfaces for the 15 and 20 layers enhance LDI. However, ion intensities for insulin b chain decreased by 47% and 44% on 15-layer and 20-layer surfaces as compared to that on the 5-layer substrate. This observation suggests a rather complex process for peptide ionization on the calcinated surface and other surface properties such as hydrophobicity may play an important role. Insulin b chain is known to be more hydrophobic than [Sar1, Thr8]-angiotensin II since insulin b chain shows a stronger retention in reversed-phase LC than [Sar1, Thr8]-angiotensin II (data not shown). The relative higher hydrophobicity of insulin b chain may cause poor dispersion of the molecules on a hydrophilic surface, especially a porous film. By contrast, [Sar1, Thr8]-angiotensin II, which is a hydrophilic peptide, tends to disperse well over the hydrophilic surface where higher abundant analyte ions were resulted. The “match of hydrophilicity” allows the molecules to penetrate effectively into pores on the surface, and therefore the efficiency of heat transfer from substrate to analytes is enhanced. Clearly the surface wetting property is important to manipulate LDI on a calcinated film, which favors samples with better dispersion on the surface.

Figure 7.

Contact angle measurement for calcinated silicate surfaces. (a) Contact angle of water on surfaces changes with layers of silicate. (b) Images of water droplets on surfaces with different layers of silicate. Error bars (< 1°) are omitted for simplicity (n>3).

The overall mechanism of LDI on a nanoscale calcinated film on Au could be complex and likely an electron-phonon collision/lattice heating phenomenon. The application of pulsed UV-laser onto the nanometer-scale Au layer leads to rapid thermalization of excited electrons, giving rise to a hot free electron gas that heats up the metal lattice through a collision mechanism^{37, 38} or volume plasmon process³⁹. We believe the calcinated film on Au plays a crucial role of confining the heat within a local area due to its low heat conductivity. The localized heating promotes vaporization of the molecules and thus desorption of analytes. Porosity of the film and match of wetting property that leads to analyte penetration into the porous calcinated film are important. The LDI process may also be assisted by surface nanostructures of the calcinated layer including small islands and sharp tips, at which ion exchange and charge separation may be involved to produce analyte ions. It should be noted that thicker films (>60 layers) attenuated the ion intensity, suggesting a delicately balanced role of the calcinated film between local confinement of heat and total insulation. The use of citric acid highly improves LDI performance for its role as an external proton donor and possibly as a “buffer” in heat transfer to enhance the “soft” desorption ionization for analyte ions.

CONCLUSIONS

We have demonstrated that nanoscale calcinated films on Au are a highly attractive and promising substrate for SALDI-MS analysis of biomolecules including amino acids, peptides and small proteins. Low background noise and high LDI efficiency offers a new platform for mass spectrometric analysis with a large mass range. The calcinated silicate substrate has several advantages over other existing SALDI-substrates, including ease of fabrication and modification, high reusability, good reproducibility, long-term air stability, and low cost. The LDI on calcinated substrates appears to depend on laser induced thermal desorption, in which the thin Au layer plays a crucial role for energy absorption and heating whereas nanoscale silicate film is important for heat confinement and generation of hot spots. Surface hydrophilicity and roughness of the calcinated film are important factors that manipulate the performance. Existence of low concentration of citric acid in sample highly promotes protonation of analytes and suppresses ion fragmentation. As surface properties of glass can be easily manipulated by silane-based chemistry and the thin gold film is optically active, the LDI-MS with the calcinated substrates are amenable for integration with existing technologies such as microfluidics, microarray chips and many optical methods.