Development of a photoacoustic acquisition system and their proof-of-concept for hemoglobin detection

Contents lists available at ScienceDirect Heliyon journal homepage: www.cell.com/heliyon Research article Development of a photoacoustic acquisition system and their proof-of-concept for hemoglobin detection Bruna Pinheiroa,b, Vânia Pintoa,b, Hugo Dinisa,b, Michael Belsley c, Susana Catarinoa,b, Graça Minasa,b, Paulo Sousaa,b, ,∗ aCenter for MicroElectromechanical Systems (CMEMS), University of Minho, Guimarães, 4800-058, Portugal bLABBELS — Associate Laboratory in Biotechnology and Bioengineering and Microelectromechanical Systems, University of Minho, Braga, 4710-057, Portugal cCentre of Physics of Minho and Porto Universities (CF-UM-UP), Laboratory for Materials and Emergent Technologies (LAPMET), University of Minho, Braga, 4710-057, Portugal A R T I C L E I N F O A B S T R A C T Keywords: Hemoglobin Laser Organ-on-a-chip Photoacoustic Piezoelectric Ultrasound Recently, Organ-on-a-Chip (OoC) platforms have arisen as an increasingly relevant experimental tool for successfully replicating human physiology and disease. However, there is a lack of a standard technology to monitor the OoC parameters, especially in a non-invasive and label-free way. Photoacoustic (PA) systems can be considered an alternative and accurate assessment method for OoC platforms. PA systems combine an illumination source to excite the sample molecules, with an ultrasound sensor to measure the generated ultrasonic waves, combining the advantages of optics and acoustic methodologies to safely acquire tridimensional signals and images at various depths. This work is focused on the design, implementation and test of an acquisition electronics circuit, based on the PA principle, for hemoglobin (Hb) detection, aiming towards a future integration within an OoC platform. Based on the measured frequency response of commercial piezoelectric transducers, an electronics design comprising a differential charge amplifier and a band-pass filter was developed. Experimentally it was verified Hb detection for concentrations of Hb between 2.5 and 10 mg/mL in aqueous solutions, roughly 48 times lower than the typical in vivo blood concentrations. This creates the possibility of developing this technique to monitor Hb at low concentrations in small volumes, which is highly appropriate for OoC devices. 1. Introduction An OoC is a system that contains cells, tissues or organ models, natural or artificial, grown inside microfluidic channels. The main purpose of an OoC is the simulation of the physiological environments of human tissues and organs, and one of its greatest potential applications is to assess the safety of drugs before they enter clinical trials, increasing the efficiency of pharmaceutical development. *Corresponding author at: Center for MicroElectromechanical Systems (CMEMS), University of Minho, Guimarães, 4800-058, Portugal. E-mail address: [email protected] (P. Sousa). https://doi.org/10.1016/j.heliyon.2024.e41083 Received 9 September 2024; Received in revised form 12 November 2024; Accepted 8 December 2024 Heliyon 11 (2025) e41083 Available online 12 December 2024 2405-8440/© 2024 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license ( http://creativecommons.org/licenses/by-nc-nd/4.0/ ). B. Pinheiro, V. Pinto, H. Dinis et al. Fig. 1. Schematic representation of the photoacoustic imaging principle. (a) Laser beam excitation (b) Increase of the local temperature (c) Rise and propagation of the pressure wave (d) Image reconstruction from the pressure waves (adapted from [10]). Therefore, these chips are designed to control the microenvironment of cells while maintaining organ/tissue functioning. By being able to combine advances in tissue engineering with nano and microfabrication, these OoC platforms can be considered the next-level systems for studying both human pathophysiology and therapeutics effects [1,2]. Hemoglobin is one of the molecules essential to monitor on these platforms. Hb is an oxygen-transport metalloprotein in the red blood cells (erythrocytes) of almost all mammals, which plays a significant role in oxygen-carrying processes and is also a typical biomarker for certain diseases [3,4]. This biomolecule, found in the red blood cells of nearly all mammals, is a metalloprotein responsible for carrying oxygen and is a typical biomarker for specific illnesses. Normal levels of hemoglobin in healthy tissues are around 120 to 160 mg/mL in blood cells and below 0.04 mg/mL in human serum [3]. Specific diseases such as heart disease, leukemia, anemia, and others can be closely linked to abnormal levels of hemoglobin in the blood [5]. Currently, several methods such as electrochemical assay, luminescence, chromatography, mass spectrometry and enzyme-linked immunosorbent assay, have been developed for Hb monitoring. However, these methods are expensive and involve multiple preconditioning steps and are not suitable for integration in OoC platforms. In addition, these methods have been limited by the low sensitivity and specificity, requiring bulky and expensive equipment [6,7]. Thus, the development of low cost, non-invasive, highly specific and label-free detection methodologies, able to be integrated into OoC platforms, is crucial. Nonetheless, a persistent challenge with OoC platforms remains the absence of an integrated and reliable monitoring method, crucial for providing real-time information on Hb distribution. PA systems can be considered as an alternative and accurate monitoring method in OoC platforms. The measurement of blood’s Hb through photoacoustic methods has been clearly validated in the literature [8]. Hb has its dominant optical absorption coefficient at wavelengths below 1000 nm, making it clearly distinguishable from other molecules, such as water and lipids, and allowing it to obtain a photoacoustic contrast for this analyte. The specificity of photoacoustic systems is obtained by the right choice of the excitation wavelength, assuring maximum absorption peaks distinct from other biomolecules. Regarding Hb, the 532 nm excitation wavelength is often considered for imaging Hb with high specificity and a strong photoacoustic effect. PA systems are hybrid systems that employ an illumination source to excite the sample, producing ultrasonic waves detected by an ultrasound transducer. This non-invasive and label-free biomedical imaging modality leverages the optimal features of both optics and acoustics, enabling the safe acquisition of one-dimensional signals and images across various scales (including organelles, tissues and organs), and with no need for ionizing radiation. Consequently, its primary benefit is the ability to gather information from the sample, as functional and molecular related, without disrupting the regular growth of cells within the OoC. The PA effect occurs when a tissue surface is exposed to electromagnetic waves, in the intensity-modulated continuous or pulsed form [9]. Following this exposure, the light penetrates the tissue target up to a specific depth, according to its wavelength. The absorption of photons, followed by chromophores’ nonradiative relaxation, leads to a swift rise in the temperature, causing the absorbing object to undergo thermoelastic expansion. The abrupt increase in pressure causes the generation and propagation of a sound wave, detectable by traditional ultrasound transducers (Fig. 1). Locating the sources of the pressure waves enables the identification of the specific position of the absorbed light by the sample or tissue, which provides crucial molecular and functional details of the sample. This work presents the development of a first prototype of the PA detection module (including sensors and instrumentation) and its experimental validation for hemoglobin detection. It is a proof-of-concept, as it aims to be the first step in the development of a novel PA device, suitable to future integration into an OoC. Heliyon 11 (2025) e41083 2 B. Pinheiro, V. Pinto, H. Dinis et al. For that purpose, a theoretical, electrical, and experimental characterization of a commercial lead zirconate titanate (PZT) transducer for the PA detection will be conducted. Piezoelectric transducers are widely used in literature for their well-established fabrication technology and scalable sensitivity [11], making them a simple and cheap option for our proof-of-concept. 2. Theory To properly excite the sample and efficiently generate broadband PA signals, the duration of the light pulse should be shorter than the times required for thermal and stress relaxation. The thermal relaxation time, which describes the thermal diffusion within a given area, can be approximated by 𝜏𝑡ℎ =𝑑2 𝑐∕𝛼𝑡ℎ, where 𝛼𝑡ℎ represents the thermal diffusivity, while 𝑑𝑐denotes the heated area characteristic dimension or the desired spatial resolution. Additionally, the stress relaxation time characterizes how the pressure waves propagate through the domain, being defined by the 𝜏𝑠=𝑑𝑐∕𝑐, where 𝑐represents the speed of sound [9,12]. Generally, pulsed lasers with nanosecond or picosecond peak duration are used as light sources for photoacoustic (PA) imaging to meet the requirements of both thermal and stress confinements [14]. Following excitation, the expansion of fractional volume, which is given by 𝑑𝑉 ∕𝑉, can be expressed as 𝑑𝑉 𝑉=−𝜅𝑝 +𝛽𝑇 (1) where 𝜅represents the isothermal compressibility of the medium, 𝛽represents the thermal coefficient of volume expansion, 𝑝represents pressure while 𝑇denotes the temperature. Assuming that both thermal and stress induced expansion are confined, it can be assumed that the fractional volume expansion can be neglected, leading to an immediate accumulation of pressure in the heated area. The initial rise of pressure, 𝑝0, is given by Equation (1) 𝑝0=𝛽𝑇 𝜅(2) or described as 𝑝0=𝛽 𝜅 𝜂𝑡ℎ𝐴𝑒 𝜌𝐶𝑣 (3) where 𝐴𝑒denotes the specified optical absorption, 𝜂𝑡ℎ relates to the optical energy absorbed and converted into heat (in percentage), 𝜌represents the density of the medium while 𝐶𝑣represents the specific heat capacity at a constant volume. The Gruneisen coefficient, which is a dimensionless thermodynamic constant, defines the efficiency of the heat-pressure conversion, allowing to simplify Equation (3) Γ= 𝛽 𝜅𝜌𝐶𝑣 (4) By replacing the Gruneisen coefficient Γin Equation (3), it becomes 𝑝0=Γ𝜂𝑡ℎ𝐴𝑒(5) In the specific conditions of linear optical absorption, i.e., when the local optical fluence 𝐹and 𝐴𝑒are proportional, 𝑝0can be given by: 𝑝0=Γ𝜂𝑡ℎ𝜇𝑎𝐹(6) where 𝜇𝑎is the coefficient of optical absorption. Pressure within the heated area immediately rises as a result of the laser’s swift energy delivery. This pressure release (as a consequence of thermoelastic expansion), leads to the production of ultrasound waves, when both the thermal linearity and confinement conditions are met. The following wave equation governs the acoustic pressure of the ultrasonic wave [9,13,15] (▽2−1 𝑐2 𝜕2 𝜕𝑡2)𝑝=− 𝛽 𝐶𝑝 𝜕𝐻 𝜕𝑡 (7) where, 𝐶𝑝represents specific heat capacity, 𝐻is the heating function (given by the heat deposited per volume and time units), and is related to 𝐴𝑒through 𝐻=𝜂𝑡ℎ 𝜕𝐴𝑒 𝜕𝑡 [9,15]. Additionally, one of the steps in this work was to characterize different piezoelectric transducers to understand their behavior. Therefore, one of the most important parameters under study is the resonance frequency of the piezoelectric transducer. The resonance frequency (𝑓𝑟) refers to the natural frequency of vibration or oscillation of an object, where the amplitude or intensity is maximum (and with lower energy losses) [16]. This frequency influences the efficiency of the mechanical wave transmission and is reliant on the transducer’s thickness, being given by 𝑓𝑟=𝑛(𝑐∕2𝑡𝑚)(8) where 𝑐is the speed of sound in the medium, 𝑡𝑚is its thickness, and 𝑛describes an odd integer that represents which odd multiple of a half wavelength is being considered, where 𝑛=1 represents the resonance frequency of the material. Heliyon 11 (2025) e41083 3 B. Pinheiro, V. Pinto, H. Dinis et al. Fig. 2. Illustration of the developed PA system. 3. Methods 3.1. Selection of the piezoelectric transducer For proceeding for the transducers’ characterization, two commercial piezoelectric disks of lead zirconate titanate (PZT) with approximately 27 mm of diameter and 500 μm of thickness were chosen [17]. 3.2. Setup for the transducer characterization In order to perform the electrical characterization of the piezoelectric transducers, a pre-calibrated Vector Network Analyzer (VNA) from Keysight (E5071C) was used. For each frequency in the kHz range, from 10 kHz up to 5 MHz, the transducers’ return loss (RL) was evaluated. Each of the PZT-5H transducers has two metallic contacts which were coated with an acrylic resin (Plastik70 Kontakt Chemie) to block the flow of current through the water medium where the transducers were immersed. Additionally, a pair of wires assures the connection to the testing equipments. Experimental characterization was carried out to investigate the correlation between the intensity of the emitted and received waves, and to analyze how variations in frequency and the distance between the two transducers affect the received amplitude. Therefore, two piezoelectric disks were placed underwater in a glass container (10 × 8 × 7 cm), enabling a better transmission of the produced mechanical vibrations. Then, a signal generator from Multicomp (MP750065) and a digital oscilloscope (Rigol, DS1102E) were employed to generate the waves and capture the emitted and received waves, respectively. More detail about the setup used for the electrical and experimental characterization of the piezoelectric transducers can be found in [17]. 3.3. Experimental setup for the PA system test To validate the PA system, several concentrations of Hb solutions (in water) were employed. The tests were performed with human Hb obtained from Sigma-Aldrich (H7379, lyophilized powder), since the main target of the application is monitoring human cells in OoC platforms. Using this solution allows us to keep the low cost of the tests and ensure reproducibility. In order to make PA signal tracking easier, the oscilloscope was synchronized with a signal from a photodiode (trigger), indicating the instant of the laser’s light pulse emission. Subsequently, a goblet containing a hemoglobin solution was positioned in the laser’s path. The piezoelectric disk, connected to the remaining detection circuit (detailed in section 4.1.2), was placed within the solution with an angle that avoids the laser hitting (approximately 45º). Fig. 2illustrates the schematic representation of the system setup. The tests were then conducted and the acoustic response was recorded. 4. Photoacoustic system design and implementation 4.1. System design 4.1.1. Illumination source To excite the sample under study, an illumination source with a wavelength in the 400-600 nm range (hemoglobin has maximum absorption at 414, 541 and 576 nm wavelengths [18]), a narrow pulse width, and an energy level of around mJ was required for our PA system. Consequently, a solid-state laser was selected. The Nd:YAG laser (LPY604T-10, Litron Lasers) (Figs. 3(a) and (b)) emitted light pulses into the PA system, with a wavelength of 532 nm, 10 Hz PRF, 20 ns pulse width, and 1.2 mJ of energy per pulse, focusing on an area of 60 μm in diameter. Heliyon 11 (2025) e41083 4 B. Pinheiro, V. Pinto, H. Dinis et al. Fig. 3. Solid-state laser used as illumination source with its optical path illustrated from different perspectives. 4.1.2. Detection module After selecting the most suitable piezoelectric transducers for detecting the PA signals in the reported system, the design of the electronics detection circuit was performed. The detection module of the PA system uses a PZT-5H piezoelectric transducer with a 27 mm diameter and 500 μm thickness. Its experimental resonance frequencies were around 94 kHz in the radial mode and 4 MHz in the TE mode (section 5.1), characterized using the setup presented in section 3.2. This transducer was chosen because it is widely used in similar systems [11,19–21] and its favorable electrical characterization, showing good impedance matching and also a linear relationship between emitted and received amplitudes in the experimental characterization. Following the piezoelectric transducer, a charge amplifier comprising a two-stage differential charge amplifier [18] was used. The first stage was based on a modified Howland circuit to convert differential-mode input to a single-ended signal while rejecting common-mode input. The second stage consisted of a standard current integrator with high-pass filtering outputted a single-ended voltage. This setup offers low noise and high sensitivity, enhancing the signal to noise ratio (SNR). Gain adjustment was achieved by regulating resistors 𝑅𝑎and 𝑅𝑏[22] (in this work, a gain of 46 dB was used). Grounded resistors were connected to each amplifier input in the first stage to supply the necessary bias current. This amplifier was implemented using the IC LT1259 (Analog Devices), supplied by a 5V DC power supply. Matching conditions for proper amplifier operation are detailed in Equation (9) and Fig. 4(a). ⎧ ⎪ ⎨ ⎪ ⎩ 𝑅1=𝑅 𝑅2=𝑅𝑎+𝑅𝑏=𝑅 (9) with ⎧ ⎪ ⎨ ⎪ ⎩ 𝑅𝑎=(1−𝑎)𝑅 𝑅𝑏=𝑎𝑅 0<𝑎≤1 For this charge amplifier, its differential gain (𝐺𝐷) and sensitivity (𝑆𝑄) are given according to Equations (10) and (11): 𝐺𝐷=2 𝑎 𝑠𝑅𝑓 1+𝑠𝑅𝑓𝐶𝑓 (10) 𝑆𝑄=2 𝑎 1 𝐶𝑓 (11) It can be verified that reducing the value of 𝑎leads to an increase in both differential gain and sensitivity. This can be achieved by increasing 𝑅𝑎and decreasing 𝑅𝑏while decreasing 𝐶𝑓also boosts sensitivity. The chosen circuit values resulted in a low 𝑎value (𝑎∼0.33). To meet low pole frequency requirements, high 𝐶𝑓and 𝑅𝑓values are recommended, but low noise and high sensitivity require a low 𝐶𝑓. Thus, a high 𝑅𝑓and low 𝐶𝑓were chosen. Secondly, a passive Twin-T notch filter was implemented to remove power-line interferences, with a 50 Hz cut-off frequency (𝑓𝑁) as projected in Equation (12) (Fig. 4(a)). 𝑓𝑁=1 4𝜋𝑅𝐶 (12) Finally, an active band-pass filter (Fig. 4(a)) was implemented with a lower cut-off frequency (𝑓𝐿) of 15.9 kHz and an upper cutoff frequency (𝑓𝐻) of approximately 5.3 MHz to attenuate very low and very high frequencies. The TL084CN IC (Texas Instruments) was chosen for the band-pass filter, and this IC was supplied with DC voltages of ±15 V. Heliyon 11 (2025) e41083 5 B. Pinheiro, V. Pinto, H. Dinis et al. Fig. 4. (a) Schematic and (b) Bode plot of the detection circuit. Both the upper and lower cut-off frequencies were calculated according to the Equation (13): 𝑓𝑐=1 2𝜋𝑅𝐶 (13) The complete detection circuit is represented in Fig. 4(a). Simulation on LTspice software analyzed the circuit’s Bode plot, as shown in Fig. 4(b). Nevertheless, it is essential to recognize that this graph represents an estimation of the circuit’s actual performance, since it was not tested with the PZT transducer. However, the obtained curve displays a significant decrease near 50 Hz, which aligns with the response of the notch filter. The increase continues until it hits a peak of 46 dB. It is important to point out that around 13.4 kHz, there is a reduction of -3 dB from the maximum gain, which corresponds closely to the lower cut-off frequency that was set for the band-pass filter. The upper cut-off frequency is around 175.1 kHz, which is lower than the designed cut-off frequency for the band-pass filter, which is attributed mainly to the Gain x Bandwidth product. Overall, the circuit has a substantial gain of 46 dB and a bandwidth of around 161.7 kHz, meeting the necessary characteristics for detecting PA signals from biomolecules (frequency around 100 kHz for hemoglobin), since even if it is necessary to detect in the Heliyon 11 (2025) e41083 6 B. Pinheiro, V. Pinto, H. Dinis et al. Fig. 5. (a) Schematic of the PCB (b) PCB connected to the PZT transducer. Table 1 Theoretical resonance frequencies calculated in different modes. Mode Thickness Resonance Frequency TE 500 um 3.98 MHz Radial 27 mm 73.3 kHz range of a few MHz, the circuit has enough gain. Furthermore, it should be noted that the gain can be easily adjusted by varying 𝑅𝑎 and 𝑅𝑏as required, with the circuit achieving a maximum gain of approximately 100 dB. The developed circuit design was implemented on a fabricated PCB to obtain a more robust and compact circuit. Fig. 5presents the design (Fig. 5(a)) and the fabricated PCB connected to the PZT transducer (Fig. 5(b)) of the developed circuit. 5. Results and discussion 5.1. Piezoelectric transducer characterization The following subsections present the results obtained from the measurements and characterization of the piezoelectric PZT-5H transducer chosen for the PA system’s detection module. 5.1.1. Theoretical characterization Firstly, it was calculated the theoretical resonance frequency of the PZT-5H disk transducer, both in the thickness and radial directions. For this calculus, it was considered the thickness and radius of the transducer, as well as the sound speed 3.98 × 103m/s of the PZT-5H material [17,23]. Hence, Equation (8) was solved, obtaining the resonance frequency values presented in Table 1. In the thickness extensional (TE) mode, the thickness value corresponds to the thickness of the transducer, whereas in the radial mode, the thickness value equals the transducer’s diameter. The predicted values show that, as expected, the TE resonant frequency is considerably superior to the radial one. Heliyon 11 (2025) e41083 7 B. Pinheiro, V. Pinto, H. Dinis et al. Fig. 6. Return loss variation according to frequency for the PZT-5H transducer from 10 kHz to 6 MHz. Table 2 The amplitude of the detected wave varies according to the amplitude of the transmitted wave. Emitting transducer amplitude (V) Receiving transducer amplitude (mV) 𝜎(mV) 0.100 19.867 0.377 0.200 34.533 1.050 0.500 79.733 0.377 1.000 154.667 0.943 2.000 301.333 1.886 3.000 452.000 0.000 4.000 593.333 1.886 5.000 752.000 0.000 As the transducers vibrate in a direction perpendicular to their thickness, they expand and contract in that direction during for each oscillation cycle. Oscillating in the TE mode causes greater deformation and requires less energy, ultimately leading to a higher stiffness when compared to the radial mode. Because the radial mode requires greater energy and has lower stiffness, the resonance frequency is lower in this orientation than in the thickness orientation [17]. 5.1.2. Electrical characterization The electrical characterization of the PZT-5H piezoelectric transducer was performed by measuring the S-parameters from 10 kHz to 6 MHz, without using an impedance matching circuit, in order to determine the return loss (RL), which indicates the reflected electrical power. During the measurements, the transducer was immersed in water. Since water and PZT have similar acoustic impedance values (𝑍=1.48 × 106𝐾𝑔∕𝑚2𝑠and 𝑍=31×10 6𝐾𝑔∕𝑚2𝑠, respectively) compared to air (𝑍= 400 𝐾𝑔∕𝑚2𝑠), this results in improved acoustic matching and reduced reflection of acoustic waves at the transducer-medium interface, allowing for more efficient sound wave propagation [17]. In Fig. 6, the RL spectrum of the PZT-5H transducer is presented, regarding to the frequency. Therefore, it was demonstrated that the PZT-5H transducer exhibits significantly lower RL peak levels at 94 kHz (approximately -22 dB) and 4 MHz (around -0.5 dB), corresponding to the radial and TE resonance frequencies, respectively. The values closely resemble the theoretical values previously discussed and represent the highest electrical transmission capability of the PZT transducer. Moreover, it can be noted that the transducer exhibits higher efficiency at the radial resonance frequency as opposed to the TE resonance frequency, as the RL value is much lower at 94 kHz. It is also noteworthy that the peaks occurring after the radial resonance frequency represent the harmonics of the signal. 5.1.3. Experimental characterization Sine waves were used during the experimental characterization to serve as input signals for the emitter transducer. The transducers were evaluated at the radial resonance frequency, which is the frequency at which the RL value is the lowest. Amplitude variations from 10 mV to 5 V were applied to the emitted wave, with the receiver placed 2 cm away. Table 2and Fig. 7illustrate the nearly linear connection between the emitted and received wave amplitudes. For instance, when the emitted wave has an amplitude of 1V, the received wave’s amplitude is around 160 mV. Then, the transducers were shifted to determine how the wave’s amplitude changes with the distance between them. Sine waves were produced at amplitudes spanning from 100 mV to 5 V, at the radial resonance frequency, and at distances of 1, 2, 4, and 6 cm. In Fig. 8and Table 3, it can be seen that the intensity of the received wave decreases as the gap between transducers increases (with variable emitting wave amplitudes), with a more pronounced decline at shorter distances that stabilizes as the distance from the transducer increases. Accordingly, the amplitude variations based on distance were deemed not significant due to the relatively low attenuation coefficient [17]. Heliyon 11 (2025) e41083 8 B. Pinheiro, V. Pinto, H. Dinis et al. Fig. 7. The amplitude of the received wave varies depending on the amplitude of the emitted wave. Table 3 Variation of the received wave amplitude as a function of the distance between transducers, for different emitting wave amplitudes. Distance E. amplitude 100 mV 𝜎(mV) 200 mV 𝜎(mV) 500 mV 𝜎(mV) 700 mV 𝜎(mV) 1 20.400 0.400 38.400 0.400 91.600 0.400 128.000 0.000 2 19.867 0.377 34.533 1.050 79.733 0.377 109.333 0.471 4 17.867 0.680 31.600 0.327 73.867 0.377 100.400 1.178 6 18.000 0.653 31.533 0.411 74.867 0.340 102.667 0.943 *E. amplitude: Emitting amplitude. Fig. 8. The amplitude of the received wave varies depending on the distance separating the PZT disks, for various emitted wave amplitudes. 5.2. Hemoglobin photoacoustic detection Fig. 9displays the mean of ten measurements of the PA signals from a sample of distilled water and of three hemoglobin concentrations. The initial peaks generated for each sample are present due to electrical interference when the laser is activated, as shown in Fig. 9, i.e., even when the laser is focused on a distilled water sample, the initial peak is also present, even though water does not produce a PA response (at this wavelength). Thus, for all samples, this first peak was neglected when evaluating the PA signal. By examining the curves of Figs. 9and 10, it was noted that more concentrated hemoglobin solutions yielded stronger responses, consistent with theory. Therefore, a maximum amplitude of approximately 44 mV was obtained for the 10 mg/mL solution. Furthermore, as shown in Fig. 10, it is observed a practically linear relationship between amplitude and hemoglobin concentration. Heliyon 11 (2025) e41083 9