Detection of airborne nanoparticles with mechanical systems

The industrial use of engineered nanoparticles has increased dramatically in recent years, raising the risk of human exposure.¹ Airborne nanoparticles can cause severe harm when inhaled, with health risks including pulmonary inflammation, asthma, chronic obstructive pulmonary disease and lung cancer.² As a result, small and cheap devices capable of detecting these particles are highly in demand by the authorities and the nanoengineering industry.³ We propose to use nanomechanical resonators to build the next-generation portable airborne nanoparticle sensors.

Nanomechanical resonators are nanometer-sized mechanical structures that vibrate at specific frequencies, like a guitar string that resonates at a defined note. These devices can detect the mass of single molecules or nanoparticles by measuring the frequency detuning when nanoparticles land on the sensor.⁴ The probability of such a particle landing on a nano-sized resonator purely by diffusion is low, which makes it difficult to collect enough in a reasonable amount of time.⁵ We overcome the low collection efficiency of diffusion-based sampling with an inertial sampling scheme: see Figure 1.⁶

Figure 1. (a) Nanoparticle sampling: the aerosol with the nanoparticles flows through a small hole in the sensor chip passing the nanomechanical string resonator. (b) Collection mechanisms: the nanoparticles are collected on the string either by inertial impaction (with efficiency E_I) or diffusion (with efficiency E_D). In the diffusive-sampling regime, thermal wiggling of the particles increases the probability of adsorption on the string when passing by on an intercepting streamline. In the inertial-sampling regime, the nanoparticles cannot follow the air streamlines around the string and impinge on it. By increasing the nanoparticle velocity u, the dominant collection mechanism changes from diffusion to more efficient inertial impaction.

In our device, the nanoparticle air stream flows through a small hole in a sensor chip containing the nanomechanical string resonator, where the particles are collected by diffusion or inertial impaction: see Figure 1. The string-like resonator essentially acts as a single filter fiber, filtering nanoparticles from the air stream with efficiency E defined as the ratio of particles captured to particles flowing in the projection of the nanostring. The particle collection by diffusion (D) is most efficient at low air velocities (u) since E_D∝u^−2/3. The efficiency E_I of inertial impaction, on the other hand, increases linearly with velocity. Therefore, the number of particles collected per second, given by the number of particles passing the resonator each second (∝u) multiplied by E_I, increases quadratically with air velocity. By forcing the air through the small orifice in the chip, we obtained velocities of more than 100m/s for low flow rates (a few 100mL per minute). Such flows can be obtained with small pumps, which is important for the design of a portable sensor. With inertial sampling, we showed that it is possible to reach a saturated sampling regime (E_I = 100%) at velocities u > 50m/s for 28nm silica nanoparticles on a 500nm-wide resonator.

We fabricated silicon-nitride nano and micromechanical string resonators (see Figure 2) and used an optical vibrometer to detect their vibration. We then coated the strings with a thin metal film to actively oscillate them at their resonance frequency by means of the Lorentz force. Finally, we used these resonators to detect nanoparticle concentration and the mass of single nanoparticles: see Figure 3.

Figure 2. Scanning electron microscope pictures of (a) a 500nm-wide silicon-nitride string resonator before measurement and (b) a 3μm-wide resonator after a measurement of 28nm silica nanoparticles. The string surface is homogeneously covered with nanoparticles (see inset).

Figure 3. (a) Detecting the concentration of 44nm sucrose nanoparticles with a 3μm-wide micromechanical resonator. We measured the concentration using a diffusion-size classifier with an accuracy of ±30%. (b) Detecting single 100nm silver (Ag) nanoparticles with a 1μm-wide nanomechanical resonator.⁶ The impact of each nanoparticle causes specific drops in the resonance frequency of the resonator and the magnitude of the individual frequency jumps is a function of particle mass and landing position along the length of the string.

In our experiments, when the nanoparticles were collected homogeneously on the entire surface of the string, the total resonator mass increased steadily and the frequency decreased linearly for constant nanoparticle concentration and air velocity. Figure 3(a) shows the oscillation frequency of a micromechanical resonator for varying nanoparticle concentrations. The oscillation frequency of the resonator drops linearly for a constant concentration and the particle concentration, can be directly obtained from the frequency slope.

When we captured a single nanoparticle on the string resonator, the resulting frequency shift Δf_n not only depended on the particle mass Δm, but also on the relative landing position X_p on the string with resonance frequency f_n (n representing the harmonic number) and total resonator mass m₀:^4,7

We measured the changes in the first harmonic resonance frequency (n=1) of the string resonator when single 100nm silver nanoparticles of mass 6fg landed on it: see Figure 3(b). Equation (1) shows that such a particle would cause a maximal frequency drop of 1100±400Hz in the oscillation frequency of the string. The frequency shifts in Figure 3(b) agree well with this maximal drop. Therefore, if the mass of the particle is unknown, we can determine it by measuring the largest frequency drops and using Equation (1).

To obtain a mass spectrum of disperse airborne nanoparticles, we need to detect the mass of every single nanoparticle that lands on a random position on the resonator. We can do this by measuring the detuning of both the first and second harmonic resonance frequencies for each particle that lands on the string (dual-mode operation), hence obtaining two frequency shifts that satisfy Equation (1) for n=1 and n=2. We can then solve the second order system of equations obtained in this way for the two unknowns: particle position and mass.

In summary, we sampled airborne nanoparticles on a nanomechanical resonator with unprecedented efficiency by streaming the air directly around the resonator at high velocity, which allowed us to measure nanoparticle concentrations and the mass of individual nanoparticles. With our highly efficient inertial sampling technique, we believe we can build a cheap and portable airborne nanoparticle mass spectrometer based on nanomechanical resonators. We are currently working on the implementation of the dual-mode operation of these mechanical systems. Because the actuation and readout are integrated, the entire sensor unit can be made very small: the largest components are a small air pump and a battery. Designing such a system is our next step. In addition to mass detection, we are working on the chemical analysis of airborne nanoparticles with the nanomechanical resonator.

We acknowledge funding from the European Community's Seventh Framework Programme (FP7/2007–2013) under grant agreement number 211464-2. S. S. acknowledges financial support from the Villum Foundation under grant agreement VKR023125.