Review Reports - Low-Cost Open-Source Biosensing System Prototype Based on a Love Wave Surface Acoustic Wave Resonator

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript presents a well-executed and timely development of a low-cost, open-source biosensing platform using Love Surface Acoustic Wave (LSAW) technology. The integration of 3D-printed components, PDMS microfluidics, and open-source electronics offers significant accessibility advantages. The study is clearly presented and technically sound. I recommend its publication in Hardware after minor revisions.

The novelty of the prototype could be more explicitly highlighted in the introduction, especially in comparison to other recent open-source SAW-based biosensing systems.
The issue of sensor degradation after repeated use, particularly the observed wear on the gold contact layer, raises concerns about long-term reproducibility and could be addressed in the discussion.
The rationale behind selecting the maximum insertion loss (IL) as the working frequency for continuous sweep measurements may require more detailed justification, particularly in the context of other possible approaches.
Inconsistent unit formatting and symbol usage can be found throughout the manuscript, such as “ul/seg” versus “µL/s”, and standardization of scientific notation would improve precision.
The section describing PDMS chip assembly and sealing discusses pressure tuning qualitatively; including more details on the determination of optimal sealing pressure would provide better insight into reproducibility.
While the conclusion touches on the system’s potential in biosensing applications, mentioning specific scenarios or analytes could help to strengthen the relevance and impact of the system.
Minor grammatical and syntactical inconsistencies appear throughout the text, including issues with article use and preposition placement; a careful language check is recommended to ensure publication-level clarity.

Author Response

RESPONSE TO REVIEWER #1

Comments 1: The novelty of the prototype could be more explicitly highlighted in the introduction, especially in comparison to other recent open-source SAW-based biosensing systems.

Response 1: We agree with this comment. The novelty is indeed presented in the introduction: the low-cost and open-source features. In addition, the development introduces an original design. To make this aspect even more explicit, the introduction text has been revised.

Text revised and marked modifications (page 2) :

Despite the advantages of LSAW sensors within the broader family of surface SAW devices, their widespread adoption has been limited by factors such as high fabrication costs, the need for specialized instrumentation and proprietary technologies that restrict accessibility [14]. A few companies manufacture both commercial SAW sensors and sensing systems based on this technology, and the available options are typically expensive and require sophisticated laboratory setups, limiting their accessibility for researchers with limited resources or for deployment in decentralized settings. To overcome these limitations, researchers often develop their own low-cost and open-source SAW-based systems, aiming to create accessible and customizable solutions [15–17]. Open-source hardware and software approaches have revolutionized various scientific fields by promoting transparency, collaboration and cost-effectiveness. In biosensing, such initiatives can accelerate innovation, enable reproducibility and facilitate technology adoption in diverse research and clinical environments.

To address the challenges explained, this work presents an original low-cost and open-source LSAW biosensing system prototype that integrates easily accessible components and customizable software.

Comments 2: The issue of sensor degradation after repeated use, particularly the observed wear on the gold contact layer, raises concerns about long-term reproducibility and could be addressed in the discussion.

Response 2: Thank you for pointing this out. The issue is addressed in Sensor reutilization and measurement technique (Page 13). The discussion on this specific point has been expanded to clarify its implications.

Text revised and marked modifications (page 13) :

While the tests did not involve surface functionalization or the use of aggressive reagents, a single sensor was used throughout all the experiments to assess how it was altered with constant use and evaluate its reutilization It was observed that the gold layer covering the four electrical contacts showed signs of degradation, with some particles detaching from the surface. This damage is likely caused by the repeated mechanical interaction between the pogo pins and the thin gold coating, combined with the pressure applied during each connection cycle. Over time, this may compromise the electrical reliability and long-term reproducibility of the measurements. To mitigate this issue, alternative contact methods with reduced mechanical stress—such as using softer contact materials or optimized pin designs—could be explored to extend sensor lifespan and maintain consistent performance. Additionally, sensor robustness could be improved by increasing the thickness of the gold layer by extending sputtering time or employing complementary gold deposition techniques such as electroplating. On the other hand, while the measurement technique used for the continuous sweep involves setting a working frequency at the maximum IL of the first sweep, other approaches can be used to explore the sensor response. One could seek the maximum IL or a specific phase value, leading to frequency shifts. Alternatively, a fixed frequency point could be chosen where IL is not necessarily at its maximum but close to zero phase, where the phase curve also exhibits linear behavior.

Comments 3: The rationale behind selecting the maximum insertion loss (IL) as the working frequency for continuous sweep measurements may require more detailed justification, particularly in the context of other possible approaches.

Response 3: We agree with this comment. In the introduction (page 2), it is explained that both IL and phase can be used to monitor changes on the sensor surface, as they exhibit similar behavior. Later, in the Custom NanoVNA Software subsection (page 5), it is mentioned that the software selects the frequency corresponding to the maximum IL as the working point for continuous sweep measurements. Finally, in the Sensor reutilization and measurement technique (page 13), other strategies may also be considered for future exploration of the sensor frequency response. We acknowledge that the reasoning presented in the design was not sufficiently elaborated, and as suggested, we have now expanded the explanation.

Text revised and marked modifications (page 5) :

The NanoVNA-type H was used to study the sensor response using its open-source software (available at: https://github.com/NanoVNA-Saver/nanovna-saver) written in Python. The VNA is a device used for various radiofrequency applications and this led to the modification of the Python code to better suit biosensing needs. The changes include removing non-essential tools, improving the interface design, and implementing real-time signal analysis algorithms, along with a continuous frequency sweep tool to assess the sensor response over a defined period. This tool sets a working frequency at the point of maximum IL amplitude and performs multiple scans over time. This approach was adopted based on the satisfactory results obtained in previous studies. For the sensor used in this work, prior research has typically selected a point near the maximum insertion loss (IL), where the phase response is expected to exhibit an approximately linear behavior[8,20,21]. In addition, various SAW-based sensing strategies in the literature analyze surface changes either by fixing a single frequency point or by comparing entire sweep curves[7,9,21,22]. Additionally, sweeps data can be exported in Touchstone format for a deep analysis, or an Excel file based on the working frequency. Figure 5 shows the customized software in operation, displaying its configuration section along with real-time graphs and data.

Comments 4: Inconsistent unit formatting and symbol usage can be found throughout the manuscript, such as “ul/seg” versus “µL/s”, and standardization of scientific notation would improve precision.

Response 4: Thank you for pointing this out. The inconsistent unit formatting and symbol usage have been corrected throughout the manuscript to ensure uniformity and precision.

Text revised and marked modifications (page 9) :

The correct operation of the system was evaluated by analyzing the baseline response of the sensor, the effect of the PDMS chip, and both fluid circulation and leakage (Figure 10). First, the frequency response of the sensor was obtained in air, without the PDMS chip, confirming that the circuit connecting the sensor to the VNA was functioning properly. Second, the PDMS chip was incorporated, and its impact on the sensor response was analyzed by varying the pressure applied to the sensor surface. Finally, distilled water was circulated with the PDMS chip tightly and loosely adjusted to observe the difference in response when fluid leakage occurred. Fluid circulation was performed with a flow rate of 0.425 µL/s.

Comments 5: The section describing PDMS chip assembly and sealing discusses pressure tuning qualitatively; including more details on the determination of optimal sealing pressure would provide better insight into reproducibility.

Response 5: We agree with this point. No specific pressure calculation was performed, as the attenuation of the IL curve was used as the indicator to ensure effective sealing between the PDMS chip and the sensor, preventing liquid leakage (page 7). A more detailed explanation of this approach has been added to the manuscript.

Text revised and marked modifications (page 7) :

A 1:10 mixture of Sylgard 184™ elastomer and curing agent is prepared and placed in a vacuum pump for one hour to remove air bubbles. Next, using the PLA mold, two metal rods, similar to those used for electronic component pins, are inserted until they reach the vertical section of the U-shaped passage (Figure 7). The mixture is then poured into the mold and placed in an oven at 52°C for at least two hours. This temperature is chosen because PLA starts to deform at temperatures above 60°C [20]. Once the PDMS is cured, the rods are removed, forming the horizontal inlet and outlet fluid channels. Finally, the chip is carefully demolded and the microfluidic device is checked for leaks by connecting hoses and circulating a simple fluid, such as water. For a simple final test, fluid circulation is performed while monitoring the insertion loss of the S21 parameter to check for potential leaks. The pressure applied to the PDMS chip involves a trade-off between ensuring proper sealing to prevent leakage and avoiding excessive attenuation that could distort the IL curve.

Comments 6: While the conclusion touches on the system’s potential in biosensing applications, mentioning specific scenarios or analytes could help to strengthen the relevance and impact of the system.

Response 6: We appreciate the suggestion. In the conclusion, we clarified that, based on the results obtained, the development is in a position to support future biosensing experiments involving the immobilization of biological elements and their testing with biofluids or specific targets. We believe that discussing particular bioreceptors and targets would exceed the scope of this work, as these choices are highly specific and depend on the particular interests and expertise of each research group. In our case, as a group that works with acoustic sensors to contribute to the diagnosis of ocular pathologies, one possible application could involve the formation of a self-assembled monolayer to immobilize an antibody functional to a given target. For example, MUC5AC has been studied for years as a potential biomarker for dry eye disease. We consider that delving into such biosensing details goes beyond the scope of this hardware-focused paper, and for this reason, we chose to only mention the biosensing potential of the system in general terms.

Comments 7: Minor grammatical and syntactical inconsistencies appear throughout the text, including issues with article use and preposition placement; a careful language check is recommended to ensure publication-level clarity.

Response 7: Thank you for highlighting the grammatical and syntactical inconsistencies presented in the manuscript. We have carefully reviewed the entire text to meet publication-level clarity.

Reviewer 2 Report

Comments and Suggestions for Authors

In this work, the authors has been development a low-cost op source LSAW biosensing system prototype was developed based on a commercially AC-quired resonator. The development integrates microfluidics through a polydimethylsiloxane (PDMS) chip, low-cost electronics and both 3D printed ultraviolet (UV) resin and pol-ylactic acid (PLA) parts. Results demonstrated that the development is able to advance to more complex applications.

This article is very well written and presented, with good grammar, and is clear and precise. The results, including tables and figures, are well-consistent and convey the robustness of the findings. Finally, the conclusion clearly summarizes the work carried out.

The manuscript can be accepted for publication in the Hardware after the following major revisions:

-Is it possible to add a specific biodetection application? It would be interesting to see the answer with a real biological sample (a couple of different natures and complexity)

- Please add a comparison table showing the advantages of this new method compared to other published methods.

-Replace the following references with more up-to-date ones: 1, 3, 4, 13 and 18.

-Review the citation format. There are some errors.

Author Response

RESPONSE TO REVIEWER #2

In this work, the authors have developed a low-cost op source LSAW biosensing system prototype based on a commercially acquired resonator. The development integrates microfluidics through a polydimethylsiloxane (PDMS) chip, low-cost electronics and both 3D printed ultraviolet (UV) resin and polylactic acid (PLA) parts. Results demonstrated that the development is able to advance to more complex applications.

The manuscript can be accepted for publication in the Hardware after the following major revisions:

Comments 1: Is it possible to add a specific biodetection application? It would be interesting to see the answer with a real biological sample (a couple of different natures and complexity)

Response 1: We appreciate the suggestion to include a specific biodetection application and the use of real biological samples. To stay within the aims and scope of the journal, the current validation focused on evaluating the hardware system. The minimal biosensing approach in validation was the use of PBS (page 12) solution for sensitivity and repeatability experiment, given its common use in biosensing applications. Incorporating a full biosensing application would shift the manuscript focus away from hardware and requires significant time for experimental design and access to appropriate resources. As a research group with several years working with SAW and QCM sensors to support the diagnosis of ocular pathologies, we cannot disclose specific results at this time. However, we can tell that biosensing experiments were performed and others are currently underway.

Comments 2: Please add a comparison table showing the advantages of this new method compared to other published methods.

Response 2: We appreciate the suggestion. This work does not propose a new sensing method and focuses on the development of a low-cost and open-source hardware platform for a commercial LSAW sensor. Instead of including a comparative table, we chose to describe the comparison in text, as many features of the developed system overlap with those found in other SAW-based sensing setups. In the literature, sensor modules and housings are often built from metal or plastic structures, and fluidic interfaces are typically implemented using PDMS chips that are either permanently bonded or mechanically mounted. These aspects were considered in order to address the reviewer comment more directly, and a new subsection entitled System comparative overview has been added to the validation section.

Text revised and marked modifications (page 13) :

5.5. System comparative overview

A comparative overview is provided to contextualize the relevance and novelty of the proposed system. This comparison highlights key differences between the developed platform and those typically reported in the literature for sensing systems based on SAW technology, as well as other commonly used acoustic platforms such as Quartz Crystal Microbalance (QCM) [8,18,20–22,24,30–37]. The system developed in this work is based on a commercial LSAW sensor, similar to some reported platforms that use either commercial or custom-fabricated devices. Despite many previous works relying on proprietary setups or only partially described designs, this system is fully open-source, promoting reproducibility and accessibility. In terms of cost, the use of affordable components and materials results in a low-cost solution, contrasting with the costly materials and instruments typically employed in other studies.

The mechanical housing is fabricated using 3D printing techniques, employing PLA and UV-resin, whereas reported systems often use metal, plastic, or 3D-printed enclosures without detailing their composition or accessibility. The microfluidic interface incorporates a removable PDMS chip, which simplifies cleaning and sensor replacement. This contrasts with other configurations in the literature, where the microfluidic component is either permanently bonded or derived from commercial microfluidics kits. In the proposed setup, fluid circulation is performed in flow mode and some works report batch operation or a combination of both. The sensor is removable and reusable, in contrast to platforms where the sensor is permanently bonded to a microfluidic chip, limiting its lifetime.

Electrical connection is achieved through a PCB interface with gold-plated pogo pins, providing a robust and solder-free solution, while alternative approaches include wire bonding or soldering, which are less practical for repeated use. The measurement instrument is a low-cost NanoVNA-H, whereas most reported systems rely on high-end, costly instrumentation. While some characteristics of the presented work overlap with existing approaches, the integration of affordability, reusability, and open-source accessibility results in a platform that offers practical advantages and distinguishes itself in the field.

Comments 3: Replace the following references with more up-to-date ones: 1, 3, 4, 13 and 18.

Response 3: We thank the reviewer's suggestion. In response, references 1, 3, 4 and 13 have been revised and replaced with more up-to-date and relevant citations. However, reference 18 has been retained, as it presents a specific PDMS chip design and a sealing mechanism based on pressure that is removable — a configuration that directly inspired the microfluidic approach adopted in our work. To further support this choice and provide a broader context, we have added additional and recent studies that employ this technique.

Comments 4: Review the citation format. There are some errors.

Response 4: Thanks for advising about the citation format. It has been carefully reviewed throughout the manuscript, and the identified errors have been corrected.

Reviewer 3 Report

Comments and Suggestions for Authors

The authors discuss the development of a low-cost, open-source biosensing system based on a commercial Love-type Surface Acoustic Wave (LSAW) resonat. Validation was carried out by assessing sensor behavior under different conditions (air, water, PBS), including fluid leakage, stabilization, and repeatability. The authors claim that the prototype can support future biosensing applications requiring phase-sensitive detection, however, the following point should be solved:

The manuscript alternates between “LSAW,” “SAW,” and “Love-type SAW.”So, standardize
Some figures (e.g., Figures 10–12) are referenced in results without accompanying detailed interpretation
Although phase shifts are used to demonstrate sensitivity, the manuscript lacks a formal LOD or a detection threshold for target analytes.
Pressure control is identified as a critical parameter, but no quantitative calibration or torque measurement procedure is described.
No direct comparison is made to existing commercial or academic systems, despite references to OpenQCM and others.
Although a two-layer enclosure is included, thermal drift or temperature sensitivity is not evaluated
Assembly instructions are detailed, but no information is provided on whether multiple systems were built to test reproducibility
Some schematic figures (e.g., Figure 2, Figure 3) have unclear annotations and lack sufficient resolution to distinguish structural features.
Standard deviations are reported in Table 2, but no statistical test is applied to confirm sensitivity/repeatability significance

Comments on the Quality of English Language

The manuscript reads fairly well overall, though a number of sentences could benefit from tightening, as they come across as a bit wordy or unclear. With some careful editing, the clarity and flow of the text would be much improved.

Author Response

RESPONSE TO REVIEWER #3

The authors discuss the development of a low-cost, open-source biosensing system based on a commercial Love-type Surface Acoustic Wave (LSAW) resonator. Validation was carried out by assessing sensor behavior under different conditions (air, water, PBS), including fluid leakage, stabilization, and repeatability. The authors claim that the prototype can support future biosensing applications requiring phase-sensitive detection, however, the following point should be solved:

Comments 1: The manuscript alternates between “LSAW,” “SAW,” and “Love-type SAW.” So, standardize.

Response 1: We agree with the observation. In the revised version of the manuscript, terminology has been standardized: “SAW” is used when referring to surface acoustic wave technology in general, while “LSAW” to specifically denote the Love configuration.

Comments 2: Some figures (e.g., Figures 10–12) are referenced in results without accompanying detailed interpretation.

Response 2: Thanks for raising this point. We have carefully reviewed the way the figures are referenced, as well as their corresponding captions. We believe that the current level of detail is sufficient to support the interpretation, while in others we have introduced modifications to improve clarity and provide additional context where necessary.

Example. Text revised and marked modifications (page 11) :

Stabilization was analyzed by monitoring changes in phase and IL while circulating distilled water. For the continuous sweep, the working point was set at the maximum IL of the first sweep. At the beginning of the experiment, both parameters did not stabilize. The changes were constant, small and there was no indication that the response would stabilize. If a leakage had occurred, the change would have been similar to what was observed in the previous section. After assembling the system and immediately circulating the fluid, particles may have continued depositing onto the sensor surface. Additionally, the pressure exerted by the PDMS chip on the sensor deforms the rectangular seal during system setup, and it is reasonable to assume that allowing the system to rest also contributes to stabilization. Therefore, it was decided to let the system settle with stagnant distilled water until the following day. After the relaxation period, distilled water was circulated again, and stability was reached. Figure 11 shows the temporal evolution of insertion loss and phase during the circulation of distilled water, highlighting the point at which signal stability is achieved after the relaxation period.

Comments 3: Although phase shifts are used to demonstrate sensitivity, the manuscript lacks a formal LOD or a detection threshold for target analytes.

Response 3: We thank the reviewer for this valuable suggestion. While the original manuscript focused on demonstrating the sensitivity of the system through phase shifts, we acknowledge the importance of including a formal estimation of the limit of detection (LOD). In this regard, the LOD calculation was studied and can be theoretically estimated based on the surface mass density using the standard deviation observed during the stable segments of the phase time-response curves. Initially, this calculation was not included in the manuscript to maintain focus and scope. To address the suggestion, we have now added text in Sensitivity and repeatability subsection (page 12) describing how the LOD can be estimated, and a comparison of the result with similar approaches reported in the literature.

Text revised and marked modifications (page 12) :

The Limit of Detection (LOD), defined as the smallest detectable change in signal, was estimated using the maximum phase noise (Δφₙ) of the standard deviation measured in the stable regime of the phase response. First of all, the phase variation can be combined with the theoretical phase-to-mass sensitivity of the sensor to estimate the surface mass density change (Δσ) [20]. Following established approaches [25,26], the LOD was calculated as a function Δσ, as shown in Equation 1:

LOD = (3xΔφₙ)/Δσ

During system characterization, the maximum standard deviation of 0.031° results in a LOD of 0.153 . The magnitude is acceptable in comparison with values reported where phase shift measurements are used to obtain the limit of detection in terms of surface mass density [26–29].

Comments 4: Pressure control is identified as a critical parameter, but no quantitative calibration or torque measurement procedure is described.

Response 4: We agree with this point. No specific pressure calculation was performed, as the attenuation of the IL curve was used as the indicator to ensure effective sealing between the PDMS chip and the sensor, preventing liquid leakage (page 7). A more detailed explanation of this approach has been added to the manuscript.

Text revised and marked modifications (page 7) :

Comments 5: No direct comparison is made to existing commercial or academic systems, despite references to OpenQCM and others.

Response 5: We appreciate the suggestion. This work does not propose a new sensing method, but rather focuses on the development of a low-cost and open-source hardware platform for LSAW-based sensing using a commercially available sensor. Instead of including a comparative table, we chose to describe the comparison in text, as many features of the developed system overlap with those found in other SAW-based sensing setups. In the literature, sensor modules and housings are often built from metal or plastic structures, and fluidic interfaces are typically implemented using PDMS chips that are either permanently bonded or mechanically mounted. These aspects were considered in order to address the reviewer comment more directly, and a new subsection entitled System comparative overview has been added to the validation section.

Text revised and marked modifications (page 13) :

5.5. System comparative overview

Comments 6: Although a two-layer enclosure is included, thermal drift or temperature sensitivity is not evaluated.

Response 6: We acknowledge the reviewer comment regarding the lack of explicit temperature sensitivity evaluation. It is well known that devices based on acoustic wave technology can be affected by thermal drift. In this case, after reviewing the specifications and material composition of the commercial LSAW sensor used, it was concluded that the sensor was unlikely to exhibit significant thermal sensitivity under typical laboratory conditions. Therefore, we consider it was not necessary to implement an active temperature control module, which in other studies is often used primarily just for thermal equilibration. To address this point, we have modified the Housing and Mounting Base (page 4) section to clarify that it was not needed. In addition, in Validation (page x) it was added that experiments were performed at room temperature.

Text revised and marked modifications (page 4 and 9, respectively) :

The sensor was protected from temperature fluctuations, air currents, and particle deposition by employing a two-layer enclosure. Studying the sensor structure and constituent materials, it was determined that an active temperature control was not necessary, as temperature variations were not expected to significantly affect the measurements [20]. Moreover, such control would primarily serve to acclimate the module enclosure, since direct temperature regulation at the sensor level could potentially introduce substantial signal variations.

-----------------------------------------------------------------------------------------------------------

The correct operation of the system was evaluated by analyzing the baseline response of the sensor, the effect of the PDMS chip, and both fluid circulation and leakage (Figure 10). First, the frequency response of the sensor was obtained in air, without the PDMS chip, confirming that the circuit connecting the sensor to the VNA was functioning properly. Second, the PDMS chip was incorporated, and its impact on the sensor response was analyzed by varying the pressure applied to the sensor surface. Finally, distilled water was circulated with the PDMS chip tightly and loosely adjusted to observe the difference in response when fluid leakage occurred. Fluid circulation was performed with a flow rate of 0.425 µL/s, and from this initial experiment onward, all measurements were conducted at room temperature.

Comments 7: Assembly instructions are detailed, but no information is provided on whether multiple systems were built to test reproducibility

Response 7: We appreciate the reviewer comment. The assembly process was documented to ensure that the system could be replicated by other researchers or laboratories. While only one physical system was constructed and tested, the design choices, including the use of standard components and reproducible fabrication steps, were made to maximize repeatability and facilitate future replication.

Comments 8: Some schematic figures (e.g., Figure 2, Figure 3) have unclear annotations and lack sufficient resolution to distinguish structural features.

Response 8: We thank the reviewer for this suggestion. We have carefully reviewed all figures in the manuscript, including the schematic ones. The resolution of the images has been improved and the annotations have been revised for a better clarity.

Comments 9: Standard deviations are reported in Table 2, but no statistical test is applied to confirm sensitivity/repeatability significance

Response 9: We thank the reviewer for highlighting this point. The distinction between the two fluids is clearly observable both graphically and in Table 2. In addition, we have done a t-test to provide the corresponding statistical parameters that technically support the significance of the results regarding sensitivity and repeatability. While we originally reported the p-value as p < 0.001 for simplicity, the actual calculated value is p = 4.04 × 10⁻⁴³, which further reinforces the strong statistical significance of the observed difference.

Text revised and marked modifications (page 13) :

The analysis was finished with a t-test to support the observed differentiation between fluids. A clear distinction was observed between the responses to distilled water and PBS, which was further confirmed by a statistical test yielding a p-value < 0.001, a t-statistic of 38.690, and a 95% confidence interval for the difference in means ranging from 0.607 to 0.674. These parameters demonstrate that the sensor was able to reliably differentiate between the two fluids across three repeated cycles each, resulting in two clearly distinguishable response groups.

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

Hardware

Low-Cost Open-Source Biosensing System Prototype Based on 2 a Love-Type Surface Acoustic Wave Resonator

Comments

The authors have made the modifications and corrections requested by this reviewer. In my opinion, the Manuscript is in a position to be accepted for publication in Hardware.

Reviewer 3 Report

Comments and Suggestions for Authors

The authors tried their best. Further clarifications are ok. I recommend the publication of the revised version.

A proper label is needed in Fig. 5. 4 panels but only "a, b, c" notation