1. Introduction

It is well known that light has stimulated human development for hundreds of years. Technological progress at the beginning of the 19th century allowed for the continuous improvement of devices that illuminate closed rooms and streets after dark. Following the invention of the kerosene lamp, scientists and innovators around the world continued their work to develop a new filament-based light source by experimenting with different materials. After many years of trial and error studies, the first current light bulb using a carbon filament placed in vacuum was lit. The founder of this invention was the research team led by Edison [1]. This moment is believed to be the beginning of the electric lighting era. Over the next few years, the bulb was refined in terms of shape and technical solutions. The subsequent mass production resulted in the now decades-long availability of this product. Despite the fact that in the 20th century new technologies appeared to replace traditional lighting, the incandescent bulb had the advantage of its spectral spectrum closest to sunlight. Moreover, the warm colour of the emitted light was pleasant to the human eye and did not disturb the production of hormones after long exposure, in contrast to LED and OLED devices. Recently, public awareness concerning environmental care has significantly increased. After the incandescent lamps were withdrawn from the market due to their low efficiency in relation to the demand for electricity, intensive work on a new generation light source with a spectral spectrum covering the entire visible range began. Recently we have reported observation of laser induced white emission of tungsten filament in Edison type bulb [2]. The first report on laser induced broadband luminescence in visible appeared in 2010 [3,4] since Wang and Tanner showed that lanthanide oxides under vacuum conditions and excitation with the focused beam of infrared laser diode. This phenomenon was intensively studied over the next decade. It turned out that broadband anti-Stokes LIE can be obtained from various types of materials [5–10]. Moreover, the coherence of such luminescence generated from graphene foam [11] and tungsten lamp [12] has been reported recently.

Here, the studies of intense broad band emission spectra in the visible and in particularly infrared range observed from a light bulb with a tungsten filament induced by the focused beam of infrared (IR) laser diode are reported. What is more, the mechanism of LIE of tungsten filament is proposed for the first time.

2. Materials and methods

The anti-Stokes LIE was observed for a tungsten filament located in a commercial incandescent bulb characterized by 60 W and 535 lumen supplied by Energy light. The emission spectra were recorded using an AVS-USB2000 Spectrometer (Avantes) in the visible (VIS) range and Ocean Optics NIRQuest 512-2.5 in the near infrared (NIR) range. Focused continuous work (CW) laser diodes operating under 808 nm or 975 nm were used as the excitation sources. The excitation spot diameter of the focused laser beam was measured by us in focal point on See3CAM_10CUG camera with resolution 3.75 µm/pixel and the total diameter is estimated as 175 µm (175·10⁻⁶ m) and for the power of laser diode of 1W the excitation power density was 0.7·10⁸ W/m². The diameter of the tungsten filament is 50 µm (50·10⁻⁶ m), so it is smaller than the diameter of the excitation spot. The emission spectra were corrected for the spectral setup (detector, optics and fibers) sensitivity. The rise and decay times of broadband LIE in the NIR range were measured using the Ocean Optics NIRQuest equipment. Figure 1 presents a schematic diagram of the experimental setup. Laser light was focused on the tungsten filament by lens, after that LIE was collected by optical fiber and registered by a spectrometer.

 

Fig. 1. Scheme diagram of measurement system of LIE of tungsten bulb.

Download Full Size | PDF

3. Experimental results

The spectroscopic characterization of the tungsten filament in a vacuum bulb involved the measurements of the continuous broadband emission induced by a CW laser diodes operating in the near infrared range in VIS and NIR as a function of the excitation power presented in Fig. 2. As expected, an increase in excitation power density resulted in rising emission intensity. The emission intensities were measured by two different spectrometers operating in the visible and infrared region. To compare the intensities of both bands, they were related to the maximum of the excitation line of 975 nm laser measured by both spectrometers. It should be noted that the total intensity of the broadband LIE spectrum in the NIR range is at least 5 times lower compared to the luminescence in VIS range. It is worth noting that broadband infrared emission was first reported by a Zhu et al. a few years ago [9,13,14].

 

Fig. 2. LIE spectra of tungsten bulb in visible and infrared under 808 nm (a) 975 nm (b) excitation and their power dependences (c,d).

Download Full Size | PDF

The continuous broadband emission bands observed both in visible and infrared were centred at 670 nm and 2000 nm, respectively. An analysis of the broadband spectrum in the NIR region showed that it is widened unsymmetrically and may be deconvoluted into two bands with emission maxima located at 1500 nm and 2050 nm, respectively (see Fig. S1 in Supplement 1).

The power dependence of broadband LIE intensities in the infrared and visible ranges measured for two different excitation wavelengths are plotted in Fig. 2(c) and Fig. 2(d), respectively. One can observe that after exceeding a certain excitation power threshold, the increase in emission intensity is very sharp regardless of the pumping source and measuring range. It can be seen that the course of the curve in the visible and infrared ranges can be fitted with a linear function above the excitation power threshold. However, for higher excitation powers in NIR range, the increases in intensity become smaller and are presumably related to some saturation processes. Such behaviour was previously observed by us for laser induced white emission of graphene [15]. One can note that the excitation power thresholds of LIE were dependent of excitation wavelength in VIS range and are significantly larger in comparison to the NIR range.

Further analysis included measurements of the kinetics of the observed luminescence. For this purpose, the rise and decay curves of LIE in infrared under 808 nm and 975 nm lines as a function of excitation power density were recorded and depicted in Fig. S2 in Supplement 1. Fitting the course of kinetics with the two-exponential function allowed to determine the average values of the emission rise and decay times (τ_av). The obtained results are presented in the insets of Fig. S2 in Supplement 1. It turned out that both the emission rise and decay times strongly depend on the excitation power density. In other words, the energy supplied to the tungsten filament affects the kinetics of the investigated phenomenon. The longest rise times of broadband emission were achieved when the excitation density was low. The increase of laser power results in the reduction of the rise time to the value of 95 ms and 75 ms at 2 W excitation of 808 nm and 975 nm laser diodes, respectively. Recent reports of the broadband anti-Stokes white emission recorded in the visible range have demonstrated that their build-up times are closely related to the times needed to transfer the charge from the valence to the conduction band [10,16,17]. Similar values of the broadband LIE rise times in VIS and NIR [2] indicate that they depend on the energy band gap of the analyzed material. For the emission decay times, the opposite trend could be observed. The shortest time was achieved when the tungsten filament was irradiated with 0.7 W laser beam. An increase of excitation power leads to the extension of LIE decay time. The observed trends for the rise and decay times of the LIE are consistent with the data available in the literature [2,10,16,18,19].

The influence of externally applied DC electric current on intensities of LIE of tungsten filament was measured in infrared upon 808 nm and 975 nm excitation lines and presented in Fig. 3 (the results for VIS are presented in Fig. S3 in Supplement 1).

 

Fig. 3. Power dependence of LIE generated from tungsten filament in NIR region upon 808 nm (a) and 975 nm (b) excitation lines recorded as a function of increasing voltage.

Download Full Size | PDF

One can see that the LIE intensities in VIS and NIR increase exponentially with applied voltage and can be well fitted by the relation $I({LIE} )\propto {P^{N(U )}}$, where an exponent N significantly decreases with voltage. The strongest intensity enhancements with the voltage were observed at low applied laser powers. A voltage enhancement in NIR range was much more pronounced almost by one order in comparison to VIS range. The voltage dependence is weak for highest excitation laser powers. The voltage influence is related to the Schottky effect manifesting in decrease of the work function$\; {\emptyset _{eff}} = \emptyset - {\Delta S}$, where for DC electric field the Schottky parameter is proportional to the root square of voltage ${\Delta S} \propto {V^{1/2}}$ [20,21].

4. Discussion

The broadband LIE is a point processes. A high excitation power density of laser irradiation may lead to increase of temperature at the irradiation spot of tungsten atoms and thermionic electron emission of electrons described by Richardson-Dushman relation [22] applied for modelling electron emission current J_RD from hot metallic surface:

The ionization leads to degradation of tungsten atom via emission of electrons, diabatic radiative transition to the W⁺ atom within the hybrid pair (W, W⁺) as is shown in Fig. 4. Here W⁺ characterizes the single ionized tungsten atom (in the literature W is also denoted as W I and W⁺ as W II).

The work function of tungsten is equal to 4.55 eV, therefore at least 3 photons with energy hν = 1.53 eV (808 nm) or 4 photons with energy hν = 1.27 eV (975 nm) are required to match the energy gap of ionization potential energy. It can be concluded that the observed processes are the resultant of a combination of the tungsten ionization process and emission induced by temperature. The question of temperature of tungsten filament in a spot of laser irradiation is important however needs more elaborate studies to take into account the laser cooling effects due to emission of photon and electron emission [24,25].

 

Fig. 4. Schematic diagram of tungsten energy potential curves, multiphoton absorption responsible for ionization and emission transitions for a tungsten atom pair (W, W⁺).

Download Full Size | PDF

An origin of the LIE bands for tungsten filament may be discussed using the model of mixed valence pairs successfully developed and applied for the understanding of luminescence and anomalous absorption processes in rare earth doped crystals [26]. The intervalence charge transfer (IVCT) model was applied for discussion of the laser induced broadband anti-Stokes white emission of Sr₂CeO₄ nanocrystals following multiphoton ionization [18]. It serves also for the understanding of intense white light emission induced by the laser irradiation of graphene as the sp²→ sp³ hybridized carbon atomic domains transition [15,27]. Within the IVCT model, the emission processes occur due to the diabatic transitions between the potential energy surfaces of neutral single tungsten atom W(W^I) and single ionized tungsten ion W⁺(W^II) within the diatomic pair (W,W⁺)→(W⁺,W). The distance between W atoms in tungsten metal is 2.74 Å. The potential energy curves and the respective electronic transitions for the (W, W⁺) pair are presented in Fig. 4. The energy level diagrams for the pumping and emission scheme of tungsten atom W and tungsten ion W⁺ are shown to illustrate the respective multiphoton ionization and emission transitions. The magnitudes of energy levels of neutral tungsten atom W and tungsten ion W⁺ were taken from the data reported by Kramida and Shirai [28] and the discussion of ionization potentials by Campbell-Miller and Simard [29]. The two broadband NIR emissions may be assigned to the diabatic ⁵D_3/2(W_I⁺)→⁵D₀(W_II) and ⁵D_3/2(W_II⁺)→⁵D₀(W_I) transitions, whereas in the VIS range to the second ⁵D_3/2(W_I⁺)→⁵D₀(W_II) tungsten atom transition.

The broadband LIE in the visible region may be attributed to the ⁵D_3/2(W⁺) → ⁵D₀(W) transition between a tungsten ion and a tungsten atom. It may be the reason why the intensity of broad band in the visible range is significantly more intense than that of the bands in the infrared range.

5. Conclusions

The broadband emission spectra of tungsten filament in a bulb were measured in the VIS and NIR ranges by irradiation with the focused beam of CW IR laser diodes. It was observed that the excitation energy threshold for VIS emission was much higher in comparison to NIR emission. An origin of broadband LIE was discussed within the intervalence charge transfer (IVCT) model. In this model the diabatic transitions occur within the hybrid of a single tungsten atom and a single ionized tungsten ion (W,W^-) in result of multiphoton ionization process. It was observed that the VIS emission was much more intense than NIR emission, however its excitation power threshold was several times shorter.