Portable Raman leaf-clip sensor for rapid detection of plant stress

Global food security faces challenges from population growth and climate change, demanding advancements in precision agriculture. Early diagnosis of plant stress is crucial to prevent yield loss. While optical techniques like reflectance spectroscopy are used, their limitations include similar spectral changes across various stresses and delayed marker accumulation. Raman spectroscopy offers a potential solution due to its specificity and ability for early diagnosis. This research aims to develop a portable Raman sensor for rapid, in-situ plant stress phenotyping, addressing the need for efficient and timely stress detection in field conditions. The development of such a tool would directly benefit farmers and plant scientists in optimizing crop management and breeding programs, leading to increased crop productivity and sustainability.

Existing Raman spectroscopy applications in plant science either involve collecting leaf samples for later analysis or positioning whole plants near a spectrometer. This study builds upon these methods by developing a novel leaf-clip Raman probe for direct, in vivo measurements on plants in their natural growing environment. Previous leaf-clip sensors have primarily focused on leaf absorption or fluorescence, potentially affected by leaf heterogeneity. This new leaf-clip Raman sensor aims to overcome these limitations by utilizing multiple analytes observable in Raman spectra to establish internal references, making the diagnostic less sensitive to leaf-specific variations in transmission and absorption.

A portable Raman leaf-clip sensor was designed and built, incorporating a Raman fiber probe connected to a portable Raman instrument using an 830 nm excitation laser. The 3D-printed leaf-clip maintains probe-to-leaf distance, holds the leaf steadily, blocks ambient light, and renders the device eye-safe. The sensor uses an anodized aluminum Raman probe with a focusing lens and filters. An 830 nm laser provides excitation light, and backscattered Raman light is collected and transmitted through an excitation-blocking filter. Laser power variations (75-405 mW) showed no significant effect on spectra, with 130 mW used for experiments. Reference Raman measurements were performed with a benchtop Raman spectrometer on leaf sections. For analysis, five spectra were collected per sample spot (10s integration time), with cosmic ray events removed, and spectra smoothed with a Savitzky-Golay filter. Fluorescence removal involved polynomial subtraction. Calibration of the Raman shift was verified using polystyrene. The sensor was tested on Arabidopsis thaliana, and various leafy vegetable plants (Kailan, Lettuce, Choy Sum, Pak Choi, Spinach) with different leaf characteristics. Nitrogen deficiency experiments involved growing plants in sufficient (+N) or deficient (-N) hydroponic media. Drought and heat stress experiments were also performed on Choy Sum. Data analysis included comparing in vivo leaf-clip measurements to benchtop measurements, generating histograms of Raman peak intensities, utilizing peak ratios for internal referencing, and calculating signal-to-noise ratios (SNR). Principal component analysis (PCA) was used to distinguish stress conditions.

The portable leaf-clip Raman sensor produced reproducible in vivo Raman spectra from diverse plant leaves. The 1045 cm⁻¹ peak was identified as a specific signature of nitrate, clearly distinguishing nitrogen-sufficient (+N) and nitrogen-deficient (-N) conditions in Arabidopsis and vegetable crops. The in vivo and in situ measurements using the leaf-clip sensor were comparable to those obtained with the benchtop Raman spectrometer on leaf sections. Using an adjacent peak (1067 cm⁻¹) as an internal reference improved the classification of nitrogen status, reducing variability across replicates and leaf locations. The SNR for peak ratios was comparable between the in vivo and benchtop measurements. Experiments on drought and heat-stressed Choy Sum plants showed variations in secondary metabolites detectable within 3 days, successfully distinguished by PCA. The sensor's effectiveness was demonstrated in detecting early nitrogen deficiency even before visible symptoms of leaf yellowing appeared.

The portable leaf-clip Raman sensor provides a rapid, non-invasive method for diagnosing plant stress, particularly nitrogen deficiency, in both model plants and commercially important vegetable crops. The sensor's portability and ease of use make it suitable for field applications, aiding farmers in timely interventions. The consistent results compared to benchtop measurements validate the sensor's reliability. The use of internal referencing enhances the accuracy of stress assessment by mitigating the effects of leaf heterogeneity. The successful application to drought and heat stress demonstrates the sensor's versatility in detecting various abiotic stresses. This technology holds considerable promise for improving crop management, enhancing resource-use efficiency, and promoting sustainable agriculture.

This research successfully demonstrated the use of a portable leaf-clip Raman sensor for the rapid and accurate detection of plant stress. The device's portability, ease of use, and reliable performance offer a valuable tool for precision agriculture. Future work could explore its application to a wider range of stresses and plant species, optimizing spectral analysis algorithms for improved diagnostic accuracy and developing user-friendly software for data analysis and interpretation.

While the study demonstrates the sensor's effectiveness, further validation is needed across diverse environmental conditions and plant species. The current analysis primarily focuses on nitrogen deficiency; future research should explore the application to a broader range of stress conditions and develop more comprehensive spectral libraries. Optimization of the data analysis algorithms could further enhance diagnostic accuracy and robustness.