1 INTRODUCTION

The motivation driving seed–plasma treatment research is the importance of food quality in the context of sustainable, eco-friendly agriculture. The world population is projected to increase to 10 billion by 2050 according to the Food and Agriculture Organization[1] and even without increasing the food supply, it is necessary to improve the quality of current food production. Most food begins with planting seeds and, therefore, techniques such as seed priming or seed coatings are used to improve plant growth parameters such as germination rate and uniformity and contribute to higher yields and greater plant resistance. Because the current demand of consumers for companies is to operate sustainably and ecologically, new approaches in agriculture are being adopted. These include precision farming, urban farming using hydroponics, vertical farming, genetic engineering or plant breeding, soil management, microbial farming nanoparticles and the use of biodegradable, nontoxic compounds or biologicals, products derived from microorganisms, herbs, and organic matter.[2-6–7] Hence, there is a constant interest in finding alternatives to minimize the use of resources, without degrading the environment, while simultaneously ensuring the healthy development of the seed into a mature plant. Within this framework, plasma treatments are being intensively investigated in the field of agriculture. Although many laboratories and industries see the potential for seed–plasma treatments, the lack of standardized protocols has created a situation where each group works with its own customized setup.

Therefore, the aim of this paper is to identify and analyze each seed–plasma treatment parameter that is relevant for reproducing experiments. This includes the design of the plasma device, the analysis of the seeds, and reporting of the results. The importance cannot be stressed highly enough in this cross-disciplinary field, where clear scientific communication requires more effort due to differences in terminology and working styles, and variations in protocols between fields and even within fields. There are issues with overlooking minor, but important details and working with implicit assumptions that should be challenged rather than passively accepted.

The evolution of a research topic with high-quality reporting and documentation is a gradual process that develops over years of research, as seen in plasma microbial inactivation that started in the mid-1990s and expanded to eukaryotic cells for medical applications in the early 2000s.[8] Our intention here is to aid the development of the plasma agriculture field by providing a structured approach, where the necessary variables are reported in a systematic manner to identify trends.

With this aim, the review proposes a checklist in Section 2 that captures many of the parameters to be used as a template for recording an experimental procedure. It is divided into four main parts, namely, Section 3: Plasma device; Section 4: Seed preparation; Section 5: Seed–plasma treatment; and Section 6: Seed posttreatment. Each parameter is presented in a dedicated subsection where the literature is briefly discussed, with the importance of the parameter and its practical implications. Finally, the areas in this field that require clarity are discussed in Section 7, before concluding.

2 PARAMETERS TO CONSIDER DURING PLASMA TREATMENT OF SEEDS

Seeds are the focus of this review because they are more robust substrates.[9-11] This review does not include seed disinfection of seed-borne pathogens by plasma because of existing reviews.[12] Instead, a summary of the parameters in seed–plasma treatment and how plasma may be affecting the seed and its development will be discussed. Of course, plasma disinfection and intrinsic plasma effects could both improve seed development.[13]

A plethora of variables come into play for both plasma physics and seed biology. The parameters to be considered are shown pictorially in Figure 1, and are listed in the checklist of Figure 2 in four main parts as follows: (i)

The first step is to describe the plasma device in Section 3. In this checklist, it is arbitrarily assumed that the components of the plasma reactor and its diagnostics are part of the laboratory infrastructure that do not change during an experimental campaign. Naturally, any of these parameters will change in experiments specifically designed to investigate their particular influence.

(ii)

All steps that include preparing the seed for the plasma treatment are listed in Section 4.

(iii)

The parameters of the plasma treatment itself are shown in Section 5. For the purposes of these plasma agriculture experiments, the control seeds undergo exactly the same overall procedure, except for the steps in this section.

(iv)

The seed posttreatment, such as seed handling, growth, and seed measurements, are covered in Section 6.

Pictorial template of a preliminary protocol for plasma treatment of seeds

Checklist of a proposed protocol for the plasma treatment of seeds, showing the section titles 3–6 and the parameter subsection titles (all in white text). Example parameters (black text) are discussed in the corresponding subsections

In the protocol checklist (Figure 2), each category is broken down into subsections with the corresponding names in the text. Each term is briefly discussed by comparing the methodologies in the literature to give an idea of current trends that lead to effective seed–plasma treatments. Some pitfalls are also highlighted with recommendations that can be adopted by current researchers and newcomers. This is not a straightforward task because most parameters are interdependent, making it difficult to pinpoint and isolate the cause(s) for the effects; however, it remains a worthwhile task to pursue for clarity. This preliminary protocol may help to ensure that all the relevant experimental parameters are fully recorded.

3 PLASMA DEVICE 3.1 Plasma source and reactor description

The mechanisms of plasma–seed treatments as a potential seed processing technology are reviewed by Waskow et al.,[14] and some of the corresponding plasma devices found in the literature are listed in Figure 3. Among the papers cited in the present study, the most common plasma sources include planar dielectric barrier discharges (DBDs) in various forms as shown in Figure 3(1),[11-64] followed by DBD plasma jets as shown in Figure 3(2)[10, 36, 65-74]; low-pressure plasma in a vacuum chamber as shown in Figure 3(3)[28, 43, 75-89]; and less common devices such as gliding arcs,[9, 43, 76, 90] and corona arrays as shown in Figure 3(4).[43, 91, 92]

Basic elements of plasma devices taken from the literature for plasma treatment of seeds. 1(a) Volume DBD shown with one dielectric barrier; 1(b) surface DBD where the top patterned electrodes can also be a grid or mesh. 2(a) Tubular DBD plasma jet; 2(b) CDPJs using a honeycomb array of hollow electrodes. 3(a) Microwave discharge creating a downstream plasma; 3(b) RF capacitively coupled reactor. 4(a) Gliding arc; 4(b) corona array. V represents a pulsed or AC voltage source; P represents microwave power. AC, alternate current; CDPJ, corona discharge plasma jet; DBD, dielectric barrier discharge; RF, radiofrequency

The various types of plasma sources in Figure 3 can be summarized as follows: A DBD generates plasma by time-varying high voltage (kV) between two electrodes; the dielectric barrier is to prevent arcing between electrodes that could otherwise occur following the electrical breakdown of the gas. The planar DBD may generate the plasma in the volume of gas between opposing electrodes (VDBD),[12, 17-41] or on the surface of a dielectric adjacent to the electrode edges (SDBD),[11, 42-53] or with embedded electrodes (diffuse coplanar surface barrier discharge [DCSBD]).[54-59] In a plasma jet, there is usually a gas flowing in a thin tube with a DBD excitation that can be pulsed DC, or continuous AC, or radiofrequency (RF).[10, 36, 65-69] A corona discharge plasma jet (CDPJ) uses a parallel array of plasma jets with gas shielding.[70-74] Low-pressure plasmas include microwave plasma[43, 75, 76] generated by a magnetron in the GHz range and RF plasma (capacitively coupled[28, 77-88] or inductively coupled[89]) that is usually sustained in the MHz range. A gliding arc device generates a controlled discharge propagating along two diverging electrodes in a gas flow at atmospheric pressure.[9, 43, 76, 90] In a corona, plasma is formed at the high-voltage tip of a sharp edge and forms diffuse plasma toward the ground electrode; corona arrays[43, 91, 92] have also been combined in a VDBD with a needle-to-plane matrix electrode.[13, 60-64]

These different types of plasma device can be expected to yield different effects for seed treatment in plasma agriculture, although only a few direct comparisons have been made. For example, different results for the germination and measurements of sprouts when comparing gliding arc, a corona array, a downstream microwave plasma, and a diffuse coplanar SDBD were evaluated by Sera et al.[43] They found that the SDBD required shorter exposure times due to its high plasma density relative to other devices. Similarly, another study was carried out comparing a gliding arc to a downstream microwave plasma, finding that the gliding arc had a positive or neutral effect, whereas the downstream microwave plasma had an inhibiting effect on the seedling length, accretion, and weight.[76] VDBD plasma required a much longer time than RF plasma for similar germination effects.[28]

Nevertheless, plasma treatment almost invariably uses DBDs because these sources are well suited to the atmospheric air environment of seeds, and cover a wider surface area than plasma jets. Although it is possible to have an array of multiple jets, it remains challenging to maintain treatment uniformity. One example of a custom-built SDBD discharge is shown in Figure 4.

Photograph of a plasma generated by a custom-built surface dielectric barrier discharge, 74-mm-diameter active area, using a regular array of 2-mm-wide fingers of gold-coated copper electrodes 0.1-mm thick, spaced 2-mm apart on a Kapton dielectric (0.3-mm thick) with a copper ground plane on the reverse side. Two electrodes are shown in yellow to aid the eye. The plasma is the most intense on the Kapton dielectric, around all the edges of the electrode fingers, where the electric field is the strongest. The photograph shows the difficulty of uniform plasma treatment of seeds in this dielectric barrier discharge source

The air environment is in stark contrast to the low-pressure, hazardous gas plasmas commonly used in the semiconductor industry where vacuum chambers are universally applied to isolate the plasma from atmospheric contamination.[93] Low-pressure plasmas are easier to ignite with a lower voltage, and reactive radicals have a longer mean free path to reach the seeds, but the inconvenience and expense of scaling up a vacuum chamber would disfavor low-pressure plasma treatments on an industrial scale. For air plasmas, a vacuum chamber is not strictly necessary, but a closed reactor provides a controlled environment. Nevertheless, there are many instances where plasma treatments are performed in open, or semiopen, environments with DBDs,[37, 52, 55-57] plasma jets,[69, 70, 94, 95] or gliding arcs.[9]

DBDs have several major advantages that, at the same time, can lead to problems: DBDs are simple and inexpensive to make, they can be adapted to arbitrary geometries, and their operation is usually reliable and robust. On the contrary, this convenience and flexibility has led to many individual innovative designs, and hence a major problem of comparison of experimental results between different groups. This contributes to the dilemma of reproducibility in plasma agriculture: the field has no accepted standard for DBD plasma sources. To the authors' knowledge, norms exist for plasma jets,[96, 97] but not for the many types of DBDs, except for the DCSBD of Roplass.[54-57] Note that commercial SDBD ozone generators are also used.[44, 51, 52]

Seed location: There are many possible configurations for the seed positions relative to the electrodes, so it is recommended to indicate the seed location in a schematic of the plasma source. More details are provided in Section 5.5.

3.2 Electrode and dielectric materials

The most commonly listed electrode materials are stainless steel, copper, or aluminum, and it is important to consider their thermal properties, chemical resistance to corrosion, and biocompatibility. The electrodes could be thin films deposited on dielectric, or nickel paste tracks, or metal structures such as grids or wire meshes. Electrodes can be water cooled,[50] or the thermal capacity of bulk metals could be used to absorb heat and minimize the temperature rise of the seed substrate during plasma treatment, depending on the duration of the experiment. Stainless steel is not as easily oxidized in the presence of ozone and, therefore, extends the lifetime of the electrodes and the reactor. On the other hand, copper electrodes are quickly oxidized and have to be replaced frequently. It should also be noted that copper has antimicrobial properties and therefore one should consider whether this changes the results of the plasma treatment and whether a biologically inert metal should be used instead.[98]

Electrodes susceptible to corrosion (such as copper) can be protected, for example, by a thin gold coating. However, the possibility of metallic nanoparticle contamination of the seed coat should be borne in mind.[52] Electrodes can also be protected from the plasma by a type of lacquer or by embedding them in a dielectric as in the DCSBD of Roplass.[54] Higher voltages are required for embedded electrodes because the plasma is further from the electric field concentrated near the electrode edges, although the plasma is more uniform for the same reason.

The most commonly listed dielectrics are glass, ceramic, alumina (of various qualities in terms of purity), Kapton, quartz, fiberglass, or FR4 (epoxy resin for printed circuit boards). The importance of the dielectric is to ensure reliable plasma ignition and maintenance of the discharge, especially under humid conditions, and to extend the lifetime of the plasma device. The dielectric can have an effect due to differences in dielectric properties, or plasma surface interactions.[22] To test the influence of dielectrics in seed–plasma treatments, two different dielectrics called Thernofase and Pertinax with Mylar in combination with either oxygen or nitrogen were used by Perez-Piza et al.[13, 63] to treat infected soybean seeds. There were no significant differences in the germination rate and vigor index when using the two dielectric materials, although disinfection seemed to be more effective with nitrogen or oxygen using Pertinax with Mylar in comparison to Thernofase.

In terms of damage from plasma exposure, the dielectric may be more in need of protection than the metallic electrodes. For example, a Kapton dielectric surface was activated by plasma exposure in the SDBD configuration of Figure 4, becoming strongly hydrophilic.

Moreover, the Kapton irreversibly degraded over time and became porous, as observed by a gradual whitening of the dielectric surface, accompanied by a shrinkage of the plasma discharge area. The plasma became “patchy,” and so the plasma effect on seeds was correspondingly nonuniform. This stresses the importance of a visual check on the uniformity of the plasma visible emission before and during experiments, and during long campaigns, to avoid false conclusions and ensure reproducibility. This is a necessary first step to assess the correct functioning of the studied devices before moving to more quantitative characterizations.

3.3 Electrode configuration

The shape, size, number, and arrangement of the electrodes determine the performance of a DBD. The electrodes can be arranged in various patterns such as fingers, stripes or honeycomb, and so forth, on one side of the dielectric, with a similar pattern or a full ground plane on the other side. A highly nonuniform SDBD plasma forms around the electrode edges where the electric field is the strongest, as shown in Figure 4. Hence, seeds placed randomly over the surface of an SDBD[47, 52] will generally experience a wide range of plasma conditions, possibly exacerbating the variance of seed–plasma treatment results. Furthermore, seeds may move during the plasma treatment due to ion wind, or electrostatic forces (see Section 5.5), thereby adding to the uncertainty of the effective plasma exposure. All of these factors should ideally be checked, for example, by fast imaging, and reported in an experimental description.

3.4 Plasma chemistry diagnostics

Apart from the electrical diagnostics of the plasma itself (see Section 5), many other in situ seed–plasma measurements are considered in the references cited here, with the plasma reactor sometimes being designed around the diagnostics to facilitate optical access. In order of increasing complexity, one can imagine a visual or video check of uniformity via appropriate windows[21]; a photograph of the discharge (see Figure 4) can be particularly helpful in a publication.[9, 21, 28, 31, 42, 47, 48, 60, 72, 89, 91, 92] Thermocouples,[9, 28, 89] fiber-optic probes,[22, 23, 57] and/or infrared imaging cameras can be used to determine surface and seed temperatures.[20, 57, 60, 99] The gas composition of the plasma can be monitored by a relative humidity (RH) probe[22]; gas sensors[9, 20, 51, 69] can be used for and ozone measurements; optical emission spectroscopy of the plasma can be performed for qualitative or quantitative measurements of electronically excited species[10, 22, 40, 48, 51, 59, 61, 91, 95, 100]; Fourier-transform infrared (FTIR) absorption spectroscopy[29, 49, 50, 59, 61] (measured directly within the FTIR sample compartment if possible)[26, 42, 53] can be performed for measurements of nonhomopolar molecules; and UV absorption spectroscopy[19, 20, 64, 91] and/or laser-induced fluorescence measurements of specific radicals can be performed.[18, 46, 67, 91] It is recommended to provide a diagram indicating where and how readings are taken in published studies. The more the measurements are made,[66] the more the potential for comprehension and comparison with other experiments.

4 SEED PREPARATION 4.1 Seed type

It seems reasonable that the effect of plasma on seed germination would depend on the individual plant species and its natural germination capacity, seed size, seed coat hardness, thickness of endosperm and surface morphology, and so forth.[56] Differences between dicots and monocots have also been mentioned by Zhang et al.,[78] indicating that dicots were more sensitive to plasma than monocots. However, with another cultivar, the same effect was not observed[30] even within the same wheat species.

When using the same plasma device across three separate studies, broccoli and radish behaved similarly, with a 2 min plasma treatment being optimal for inactivation of microbes and plant growth, whereas more than 1 min of plasma exposure had a negative effect on rapeseed.[70, 73, 74] Likewise, Saessal and Saechal barley seeds were compared with respect to the differences in their GABA content, showing that the optimal plasma treatment condition varied based on the seed type.[47] Plasma treatment on Arabidopsis Columbia (Col) and Landsberg erecta (Ler) seedlings was investigated by Kobayashi et al.[11] and although the growth was negatively affected in both, Col was more sensitive and lost chlorophyll, whereas Ler remained the same, suggesting that the genetic background leads to a different response. The variation in response due to ecotype has been previously pointed out by LoPorto et al.[77]

In Reference [101], seeds were soaked in water and treated with plasma, generating plasma-activated water (PAW) in situ, or forming the PAW first and soaking the seeds in it. Among wheat, sticky bean, lettuce, dianthus, tomato, mustard, radish, and mung bean, only the last three showed statistically significant results. Although PAW is different from a dry plasma treatment, this simply shows that different seed types require different plasma treatment conditions. Differences in the sensitivity to plasma treatment among three hemp cultivars (Carmagnola, Bialobzeskie, and Finola) were explored, indicating better growth for cultivar Finola.[76]

Differences between seeds might be due to the surface structure. For coffee seeds, which have thick and tough teguments, a longer treatment time of 120 s showed a better effect on germination, whereas shorter treatment times of 30–60 s were sufficient for grape seeds.[35] Corn and eggplant were treated by Sidik et al.,[94] showing optimal plasma treatment times of 3 and 5 min, respectively. In reference [58] the authors provide a list of plant species that germinate well after plasma treatment, such as Chenopodium album agg., Papaver somniferum, Zea mays, Pisum sativum, Brassica napus, Morus nigra, and Raphanus sativus, whereas Avena sativa and Rhododendron smirnowii do not. Otherwise, significant differences are found in Fagopyrum esculentum and Cannabis sativa when different apparatus are used. The reason remains unclear since the type of response may be due to genetics, seed coat structure, or chemistry such as pigments in the seed coat acting as antioxidants and contributing to seed coat hardening.[102] Nevertheless, it is clear that treatments need to be tailored to each seed type or groups of similar seed types, and will require optimization for each setup to ensure reproducibility.

4.2 Seed source, storage, age at application, and preselection

First, the source of the seeds should be stated to know whether the seeds will all be similar, or if there will be a higher natural variation within or across seed batches. For example, there will be fewer differences between seeds grown in the laboratory under controlled conditions with limited contamination,[48] compared to those bought at a local market[99] or retailer[39] which may not have been sterilized or pretreated, or those that are sold commercially where they may be pretreated. Seeds may be collected in the wild, with conditions and climate carefully noted.[77] Another source option are seed banks where seeds are preserved for an extended amount of time, which could limit the variation between generations.

If grown by the experimenters themselves, the seed source, treatment and seed age are determined by their individual choices. On the contrary, if commercial seeds are used, it is possible, in principle, for others to purchase the same seeds thus facilitating cross-comparison between studies. Regardless of the source, information such as germination rate, any pretreatment such as sterilization, along with the specific seed type and, if applicable, cultivar,[48, 76] should also be stated as this too can affect the results of the seed–plasma treatment. For example, if a particular commercial pretreatment already ensures a high germination rate of the source seeds, it is clear that a plasma treatment experiment has little chance of demonstrating a statistically significant improvement.

Storage: Not all studies report the seed storage conditions, although this is important to avoid variability of the seeds, their interaction with the plasma, and hence the final results. Typically, dry seeds are stored in the dark between 4°C and 10°C regardless of the seed type. For example, Norway spruce seeds[79] and purple cornflower seeds[103] were stored in the dark for 6 months at , while soybeans were stored in the dark at .[13] Wheat grains were stored between 0°C and 4°C[31, 104] or at in the dark.[56] Peas were stored in the dark at ,[55] while the ri