TEACHING ELECTROSTATIC FIELD
MODELLING THROUGH MATHEMATICAL
AND COMPUTATIONAL REPRESENTATIONS
ENSEÑANZA DE LA MODELIZACIÓN ELECTROSTÁTICA
MEDIANTE REPRESENTACIONES MATEMÁTICAS Y
COMPUTACIONALES
Ernesto Araujo Chavarro
CEEMA, Colombia
José Daniel Alape Díaz
CEEMA, Colombia
Cristian Andrés Lozada Monje
CEEMA, Colombia
Hoover Stiven Scarpetta Castro
CEEMA, Colombia
Manuel Alejandro Cleves Ortega
CEEMA, Colombia
Ernesto Araujo Chavarro
CEEMA, Colombia

pág. 6033
DOI: https://doi.org/10.37811/cl_rcm.v10i5.26441
Teaching Electrostatic Field Modelling Through Mathematical and
Computational Representations
Ernesto Araujo Chavarro1
e.araujo@udla.edu.co
https://orcid.org/0000-0002-7974-6969
CEEMA
José Daniel Alape Díaz
j.alape@dla.edu.co
https://orcid.org/0009-0008-3033-0111
CEEMA
Cristian Andrés Lozada Monje
cris.lozada@udla.edu.co
https://orcid.org/0009-0007-8551-0796
CEEMA
Hoover Stiven Scarpetta Castro
h.scarpetta@udla.edu.co
https://orcid.org/0009-0001-6344-9548
CEEMA
Manuel Alejandro Cleves Ortega
m.cleves@udla.edu.co
https://orcid.org/0009-0008-1000-5071
CEEMA
Ernesto Araujo Chavarro
e.araujo@udla.edu.co
https://orcid.org/0000-0002-7974-6969
CEEMA
ABSTRACT
This article examines how relationships among charges, potential, vector fields and computational
representations can inform university electrostatics teaching. A narrative documentary review focuses
on the correspondence between physical assumptions, mathematical operations, representational
resources and limits of interpretation. Disciplinary literature is connected with educational work on
electric-field teaching, multiple representations and pedagogical content knowledge. The analysis
distinguishes prescribed-charge superposition from induced-charge problems, scalar potential from
vector field, and numerical accuracy from physical and educational validation. These distinctions
support proposed prediction, comparison and justification tasks connecting analytical expressions,
potential maps, vector diagrams and simulation outputs. Computational applications serve as cases for
discussing what each observable permits one to conclude, without equating indicators of different kinds.
The contribution is an organisation of criteria for selecting and discussing representations in university
physics education. No classroom intervention, learning measurements or validated teaching sequence is
reported. Educational implications are proposals derived from documentary analysis and require
empirical assessment before effectiveness can be attributed to them. The review is selective rather than
exhaustive, and its conclusions are limited by the available documentary evidence.
Keywords: physics education, mathematical models, simulation models, higher education
1 Autor principal
Correspondencia: e.araujo@udla.edu.co

pág. 6034
Enseñanza de la Modelización Electrostática Mediante Representaciones
Matemáticas y Computacionales
RESUMEN
Este artículo examina cómo las relaciones entre cargas, potencial, campo vectorial y representaciones
computacionales pueden orientar la enseñanza universitaria de la electrostática. Se desarrolla una
revisión documental narrativa centrada en la correspondencia entre supuestos físicos, operaciones
matemáticas, recursos de representación y límites de interpretación. La literatura disciplinar se articula
con antecedentes sobre enseñanza del campo eléctrico, representaciones múltiples y conocimiento
pedagógico del contenido. El análisis distingue la superposición de cargas prescritas de los problemas
con carga inducida, el potencial escalar del campo vectorial y la precisión numérica de la validación
física y educativa. A partir de estas distinciones se proponen tareas de predicción, comparación y
justificación que relacionan expresiones analíticas, mapas de potencial, diagramas vectoriales y
resultados de simulación. Las aplicaciones computacionales se utilizan como casos para discutir qué
permite concluir cada observable, sin equiparar indicadores de naturaleza diferente. La contribución
consiste en organizar criterios para seleccionar y discutir representaciones en la formación universitaria
en física. No se presentan intervenciones de aula, mediciones de aprendizaje ni una secuencia didáctica
validada. Las implicaciones educativas constituyen propuestas derivadas del análisis documental y
requieren evaluación empírica antes de atribuirles eficacia.
Palabras clave: enseñanza de la física, modelo matemático, modelo de simulación, enseñanza superior
Artículo recibido 15 setiembre 2026
Aceptado para publicación: 21 octubre 2026

pág. 6035
Ensino da Modelagem Eletrostática Mediante Representações Matemáticas
e Computacionais
RESUMO
Este artigo examina como as relações entre cargas, potencial, campo vetorial e representações
computacionais podem orientar o ensino universitário de eletrostática. Desenvolve-se uma revisão
documental narrativa centrada na correspondência entre pressupostos físicos, operações matemáticas,
recursos de representação e limites de interpretação. A literatura disciplinar é articulada com estudos
sobre ensino do campo elétrico, representações múltiplas e conhecimento pedagógico do conteúdo. A
análise distingue a superposição de cargas prescritas dos problemas com carga induzida, o potencial
escalar do campo vetorial e a precisão numérica da validação física e educacional. Com base nessas
distinções, propõem-se tarefas de previsão, comparação e justificação que relacionam expressões
analíticas, mapas de potencial, diagramas vetoriais e resultados de simulação. As aplicações
computacionais são utilizadas como casos para discutir o que cada observável permite concluir, sem
equiparar indicadores de naturezas diferentes. A contribuição consiste em organizar critérios para
selecionar e discutir representações na formação universitária em física. Não são apresentadas
intervenções em sala de aula, medidas de aprendizagem ou uma sequência didática validada. As
implicações educacionais são propostas derivadas da análise documental e requerem avaliação empírica
antes que se possa atribuir eficácia a elas.
Palavras-chave: ensino de física, modelo matemático, modelo de simulação, ensino superior

pág. 6036
INTRODUCTION
University electrostatics teaching requires students to coordinate quantities that refer to the same
configuration but encode different information. A charge distribution specifies sources; potential assigns
a scalar value to position; the electric field assigns a vector; and a computational display selects how
some of those properties become visible. An equation, a colour map and a set of arrows can therefore
be mutually consistent without being interchangeable. The educational problem addressed here is how
to make their relationships and limitations explicit when teaching field modelling, especially when
three-dimensional configurations exceed the convenience of elementary symmetry arguments.
This problem cannot be reduced to providing more elaborate visualisations. A learner may reproduce a
potential surface without explaining the direction of its gradient, or calculate a field magnitude while
adding contributions as scalars. Conversely, a correct explanation may use a reduced representation
whose assumptions are appropriate. These are analytically constructed examples of reasoning demands,
not misconceptions measured in the present study. Research on conceptual knowledge in electricity and
magnetism provides a reason to assess understanding separately from successful equation manipulation
(Maloney et al., 2001).
The guiding question is: which distinctions among physical models, mathematical operations and
computational representations can inform the design and interpretation of university learning tasks about
electrostatic fields? The objective is to articulate disciplinary and educational literature into criteria for
selecting representations, asking explanatory questions and delimiting claims based on simulations. The
contribution concerns the structure of the knowledge to be taught and the evidence that students might
be asked to produce. It does not establish that a particular interface, activity or instructional approach
improves achievement.
Three-dimensional modelling is treated as an object of critical learning rather than an automatic indicator
of sophistication. Students need to distinguish a three-dimensional physical domain from a surface plot
of a scalar function of two coordinates. They also need to determine when symmetry permits a reduced
solution and when interfaces or finite electrodes demand a fuller description. Accordingly, the article
connects electrostatic equations to representational decisions and then uses published computational

pág. 6037
applications to examine the scope of evidence. This educational organisation changes the purpose of
comparison: the issue is what can be explained and justified, rather than which solver is universally best.
Background
The educational and disciplinary sources make different contributions. Villamizar (2025) develops a
conceptual progression from mechanical work and potential energy towards electric potential and field,
supported by geometrical representations and a numerical algorithm. Nascimento Neto and Siqueira
(2025) present visual treatments of Laplace problems that connect boundary conditions with potential
and field representations. These works offer material for examining how a mathematical operation is
linked to an interpretation. Their presentation of instructional resources does not, by itself, supply a
transferable estimate of learning gains.
Other contributions emphasise computational implementation. Tan et al. (2024) describe a MATLAB
graphical interface addressing Coulomb’s and Gauss’s laws. Montilla López et al. (2025) discuss a
computational treatment of electrostatic potential using Excel and Maple in an engineering-teaching
context. Such resources differ in the degree to which a learner can inspect the equations, modify the
source configuration or follow an iterative calculation. In this review, those features are considered
opportunities for designing questions, not evidence that one software environment produces better
learning than another.
The study by Reyes Roncancio and Martínez (2013), published in TED, examines pedagogical content
knowledge in electric-field teaching through a qualitative case involving a preservice physics teacher.
Its use of a questionnaire, interview and lesson-plan analysis provides an educational point of reference:
disciplinary knowledge becomes a teaching object through decisions about explanation and
representation. The present review does not reproduce that investigation or extend its findings to all
physics teachers; it uses the study to locate the instructional decisions that a review of electrostatic
models must address.
A more recent TED case study by Viafara-Ortiz et al. (2025) examines pedagogical content knowledge
in a novice physics teacher using interviews and classroom observations. It situates teaching choices
within a professional history rather than treating them as properties of a technological tool. Together,
these educational antecedents motivate a distinction between the availability of a representation and the

pág. 6038
teacher’s decisions about its use. The present contribution concentrates on that distinction for
electrostatic models, while recognising that documentary analysis cannot replace observations of how
teachers and students actually work.
Conceptual framework
Physical models and representational choices
A physical model specifies a domain, sources, constitutive relations and conditions at boundaries or
infinity. Its mathematical formulation identifies relations among quantities within those assumptions. A
numerical method approximates a formulation, and a visual representation displays selected outputs
using conventions such as colour, arrow length or line placement. These levels should be connected in
teaching but should not be collapsed. Changing a colour scale alters a display; changing permittivity
alters a material model; changing a boundary value can alter the physical problem itself.
The distinction matters when evaluating an explanation. An arrow may correctly indicate field direction
while conveying no field magnitude if all arrows have been normalised. A potential map may be accurate
but insufficient to determine a derivative without spatial information. A converged linear solve may
satisfy an algebraic tolerance without adequately resolving a singularity or interface. The analysis
therefore treats representation as an explicit selection of information. Its educational value depends on
which relation learners are expected to infer and which omitted information they must recover from
another representation.
Ainsworth’s (2006) DeFT framework distinguishes the design, functions and learning tasks associated
with multiple representations. It provides a conceptual reason to avoid assuming that displaying several
formats automatically supports understanding. Here, that perspective informs the organisation of
explanatory tasks: representations are paired around a specified relation, such as superposition or
differentiation, and learners are asked to explain the connection. DeFT is used as an interpretive
reference, not as an intervention whose effectiveness is tested in this article.
Kinds of evidence and educational inference
Three questions organise the interpretation of evidence. First, does the mathematical or computational
result follow from the stated assumptions? Second, does the physical model reproduce an appropriate
measured observable under specified conditions? Third, do learners develop the intended understanding

pág. 6039
when working with the model? These questions require different evidence. A residual or convergence
study addresses numerical behaviour; a comparison with measurements addresses a physical prediction;
and an educational assessment requires information about learners, tasks and the interpretation of their
responses.
The categories used below are analytical distinctions rather than a validated scale of competence. Source
specification, scalar vector coordination, boundary interpretation, dimensional reasoning, numerical
verification and evidential judgement describe aspects of an explanation that a teacher might examine.
Their use does not establish a developmental sequence or a universal hierarchy of difficulty. The
proposed teaching implications remain conditional on students’ mathematical preparation, available
time and the purpose of the course. An introductory class may prioritise superposition and potential
differences, whereas an advanced class may examine anisotropic media or weak formulations.
METHODOLOGY
The article retains a narrative documentary-review approach. Its starting material is a bounded set of 50
bibliographic records assembled for the study of electrostatic-field formulations and computational
applications. That number describes the initial documentary working set, not 50 eligible educational
studies or 50 references cited in the present article. The educational reorientation selects material
relevant to the relationship between representations and teaching, and adds explicitly identified
conceptual and educational antecedents. It does not reconstruct a systematic database search or claim
exhaustive coverage of physics education research.
Four complementary references support the educational framing: Ainsworth (2006), Maloney et al.
(2001), Reyes Roncancio and Martínez (2013), and Viafara-Ortiz et al. (2025). They were located
through targeted searches by topic and title and consultation of publisher, journal or institutional records.
Their inclusion is purposive: they address multiple representations, conceptual assessment or
pedagogical content knowledge. They are not counted retrospectively as members of the initial 50-
record working set. No search-flow totals, duplicate counts or screening agreement coefficients are
asserted.
The unit discussed is a documentary claim about a formulation, representation, computational output or
educational resource. Bibliographic identity, the physical regime and the reported observable are

pág. 6040
distinguished from the interpretation proposed in this review. Available article text and publisher
descriptions support the descriptions of the selected works; access was uneven, and no uniform full-text
appraisal of every initial record is claimed. Consequently, an inaccessible or unreported educational
evaluation cannot be treated as evidence that a resource has no educational value. It limits what this
review can conclude about its effectiveness.
The exposition compares source representation, boundary assumptions, mathematical transformations,
visual conventions and the type of evidence used to support a claim. Equations provide disciplinary
constraints for that comparison. Tables distinguish reported properties of published work from teaching
questions proposed here. The resulting synthesis is interpretive, without statistical aggregation of
learning outcomes, computational errors or execution times. There are no participants, classroom
observations, assessment scores, new experiments or newly executed simulations. The activities
described in the results are instructional proposals for later examination; they are not reports of
implemented teaching.
RESULTS AND ANALYSIS
From charge configurations to vector superposition
In an infinite, homogeneous, linear and isotropic medium, the Coulomb representation for prescribed
point charges is given by equations (1) and (2). The observation position is 𝐫, charge 𝑖 is located at 𝐫𝑖,
its value is 𝑞𝑖, and 𝜀 denotes absolute permittivity. Evaluation points exclude the singular charge
positions. The potential is expressed relative to a reference compatible with behaviour at infinity
(Carrillo Avila, 2025; Polito, 2025).
𝜙(𝐫) = 1
4𝜋𝜀 ∑ 𝑞𝑖
∥ 𝐫 − 𝐫𝑖 ∥
𝑁
𝑖=1
. (1)
𝐄(𝐫) = −∇𝜙(𝐫) = 1
4𝜋𝜀 ∑ 𝑞𝑖
𝑁
𝑖=1
𝐫 − 𝐫𝑖
∥ 𝐫 − 𝐫𝑖 ∥3 . (2)
Vector components must be summed before the norm is calculated. Adding individual field magnitudes
removes directional cancellation and gives a different result from the physical field. For a continuous
distribution, the sum is replaced by an integral involving the volume charge density 𝜌. Surface and line

pág. 6041
distributions require the corresponding density and measure. Polito examines why improper integrals
and asymptotic conditions require explicit treatment when sources extend indefinitely (Polito, 2025).
𝐄(𝐫) = 1
4𝜋𝜀 ∫ 𝜌
𝛺
(𝐫′) 𝐫 − 𝐫′
∥ 𝐫 − 𝐫′ ∥3 d3𝑟′. (3)
Becerra Becerra et al. represent finite-size bodies through sets of charge sub-elements and examine
spherical and non-spherical geometries. Their procedure retains Coulomb superposition and explores
the approximation of an extended distribution by a discrete collection (Becerra Becerra et al., 2026).
Discretising a prescribed charge distribution, however, does not itself solve induced charge
redistribution on an equipotential conductor. In that problem, surface charge is an unknown linked to
boundary conditions rather than an input that can be distributed uniformly for convenience.
Dimensional consistency requires [𝜙] = V, [𝐄] = V m−1 = N C−1, [𝜌] = C m−3 and [𝜀] = F m−1.
Coordinates expressed in millimetres must be converted before they are combined with International
System constants. Potential and field-strength plots do not share the same physical scale even when both
use colour maps. Carrillo Avila’s discussion of permittivity provides a conceptual and dimensional
check rather than a performance evaluation of a 3D algorithm (Carrillo Avila, 2025).
For teaching, the relevant operation is the transition from a physical configuration to a vector
contribution at an observation point. A task can ask students to mark source and observation positions
before computing any magnitude. They then identify the displacement vector, determine the sign of each
contribution and predict whether components reinforce or cancel. The calculation becomes evidence for
a prior geometrical argument rather than a replacement for one. This task is a proposal derived from the
structure of the equations; no classroom outcome is inferred from its mathematical validity.
Two equal positive charges at positions symmetric about the origin illustrate a useful distinction. Their
fields cancel at the midpoint, whereas their potentials add when the reference is at infinity. Replacing
one charge by an equal negative charge makes the midpoint potential zero while the fields reinforce
along the line joining the charges.
These idealised examples show why a zero value of potential does not generally imply a zero field. They
also separate cancellation of vector components from cancellation of scalar contributions without
requiring a large numerical model.

pág. 6042
A computational representation can make this comparison inspectable by presenting the signed
components and the scalar potential at the same locations. Normalised arrows should be labelled as such,
and a table of component values can accompany a colour map. The explanatory requirement is to justify
why the representations agree, including at points where a magnitude or potential vanishes. An isolated
image is insufficient evidence of that coordination.
Potential, boundaries and the meaning of spatial variation
The macroscopic formulation for a linear dielectric distinguishes the electric field from electric
displacement. The density 𝜌𝑓 denotes free charge, and 𝛆 allows anisotropy. Equations (4) and (5)
establish this distinction; inserting a position-dependent permittivity into Coulomb’s formula is not an
automatic replacement for solving a heterogeneous boundary-value problem (Ballas, 2025; Carrillo
Avila, 2025; Han et al., 2025).
∇ × 𝐄 = 𝟎, ∇ ⋅ 𝐃 = 𝜌𝑓, 𝐃 = 𝛆𝐄. (4)
−∇ ⋅ (𝛆∇𝜙) = 𝜌𝑓. (5)
For constant scalar permittivity, the equation becomes ∇2𝜙 = −𝜌𝑓/𝜀; in a free-charge-free region under
the same constitutive assumption, it reduces to Laplace’s equation. Dirichlet conditions prescribe
potential, whereas Neumann conditions prescribe its normal derivative or the associated flux. With 𝑔 =
𝐧 ⋅ 𝐃, the flux condition is −𝐧 ⋅ 𝛆∇𝜙 = 𝑔. A pure Neumann problem requires compatibility between
total flux and charge, together with a reference removing the additive indeterminacy. Cho’s study
concerns mixed boundary conditions and does not remove the need to specify their physical meaning
(Cho, 2022).
At an ideal interface without a dipole layer, potential is continuous. If 𝐧 points from medium 1 to
medium 2 and 𝜎𝑓 denotes free surface charge density, the jumps in equation (6) apply. The normal
component of 𝐄 can be discontinuous even without free surface charge when permittivity changes.
Interpolation that imposes continuity on every field component may therefore smooth a physical
discontinuity that the model must preserve.
𝐧 × (𝐄2 − 𝐄1) = 𝟎, 𝐧 ⋅ (𝐃2 − 𝐃1) = 𝜎𝑓. (6)

pág. 6043
Lima studies the field on a uniformly charged cylindrical shell using a representation by rings and elliptic
integrals (Lima, 2025). This problem requires distinguishing the two one-sided limits from a value
assigned on the ideal surface. A regularised mean value does not invalidate Gauss’s law, nor does it
imply that every surface field equals half the normal jump. Geometry, contributions from external
sources and the definition of the limit enter that identification. The review retains the result within its
geometric context rather than treating it as a universal interface condition.
The instructional connection is between a scalar distribution and its spatial rate of change. Equipotential
contours encode equal values, while the gradient depends on how those values vary with position.
Learners can compare two plots that use different vertical scales but represent the same potential data,
and explain why changing the display does not change the physical field. They can also shift every
potential value by the same constant and identify which predicted quantities remain invariant. These
comparisons make the reference dependence of potential visible without confusing it with an ambiguity
in the electric field.
Boundary conditions provide a second explanatory focus. Specifying a conductor potential is not
equivalent to prescribing a uniform surface charge on an arbitrarily shaped conductor. A classroom
comparison can keep the geometry fixed while changing the physical quantity assigned at the boundary,
then ask which unknown must be solved for. At a dielectric interface, students can annotate which field
components must satisfy continuity or jump relations and from which side a numerical value is sampled.
This connects the calculation to the physical meaning of its input rather than treating boundary settings
as software options without interpretation.
The distinction between geometric dimension and display dimension should be explicit. A graph of
potential above a two-coordinate domain has a third plotting axis for the value of the potential; that axis
is not automatically a third spatial coordinate. Conversely, a planar section of a three-dimensional
solution retains only part of its spatial information. A teaching task can request the domain coordinates,
the plotted quantity, units and any symmetry assumption before interpreting the figure. The task assesses
representational description, not an ability to produce the most visually elaborate image.

pág. 6044
Alternative formulations and their domains of validity
Salazar et al. develop a representation of surface-electrode fields through an electric vector potential and
an integral analogous to Biot–Savart (Salazar et al., 2022). The analogy is mathematical: it does not
transform electrostatic charges into magnetic currents. Its scope includes charge-free regions and
specific conditions in a planar electrode system. Conditions on the conductors must be distinguished
from those imposed in the gap; assigning Neumann conditions indiscriminately to every surface changes
the physical problem.
𝐄 = ∇ × 𝚯 ⇒ ∇ ⋅ 𝐄 = 0. (7)
Identity (7) identifies a restriction that must accompany this representation. In a homogeneous medium
with nonzero volume charge, the field cannot globally be expressed solely as the curl of a regular
potential. In heterogeneous media, the absence of free charge implies ∇ ⋅ 𝐃 = 0 but not necessarily ∇ ⋅
𝐄 = 0. Beyond this local restriction, global existence depends on the domain and its fluxes. Equivalence
between formulations must therefore be checked for the same sources, boundaries and topological
conditions; the corpus does not establish general superiority of the electric vector potential.
The educational purpose of introducing this advanced formulation is to examine the conditions attached
to an equivalence claim. Students can be asked to identify which divergence statement is required, where
it holds and whether the domain introduces global restrictions. This does not require presenting the
electric vector potential as a replacement for the scalar potential in an introductory course. It offers an
advanced comparison in which a familiar vector-calculus identity constrains the admissible physical
representation. The mathematical analogy with magnetostatics must be explained as an analogy between
integral structures, not as an identity between their sources.
Numerical methods as objects of explanation
The finite-element method approximates a weak formulation of the potential problem. For a test function
𝑣 vanishing on the Dirichlet boundary, and using the preceding flux convention, the structure in equation
(8) follows. The discrete matrix incorporates geometry and material constitution. Solving it provides
potential coefficients; the field subsequently requires differentiation or a compatible reconstruction. The
studies by Han et al. and Liao et al. illustrate applications where interfaces or three-dimensional
geometries enter explicitly (Han et al., 2025; Liao et al., 2023).

pág. 6045
∫ ∇
𝛺
𝑣 ⋅ 𝛆∇𝜙 d𝛺 = ∫ 𝑣
𝛺
𝜌𝑓 d𝛺 − ∫ 𝑣
𝛤𝑁
𝑔 d𝛤. (8)
Arun et al. investigate an algebraic topological method for three-dimensional Laplace problems (Arun
et al., 2023). Its interest lies in the discrete representation of the problem’s relationships, which must be
distinguished from metric and constitutive choices. Exact topological identities in a discretisation do
not, by themselves, imply exactness of the approximated field. Without consistent information on
system size, tolerance, hardware and error, individual comparisons cannot establish a general ranking
against finite elements.
Finite differences offer a direct implementation on structured meshes. Montilla López et al. use the
Liebmann method on a square domain and compare computational presentation in Excel and Maple
(Montilla López et al., 2025). This work supports a teaching treatment of Poisson’s equation, but a
plotted potential surface does not make it a three-dimensional validation. Similarly, Wang et al.’s study
using superLU is an explicitly two-dimensional conference contribution; superLU is the algebraic solver
rather than an alternative physical formulation (Wang et al., 2022).
Han et al. combine a direct-implicit particle-in-cell method with trilinear anisotropic immersed finite
elements (Han et al., 2025). Their model uses Cartesian meshes independent of the interface and allows
larger spatial and temporal steps in the cases examined. Stability, temporal accuracy and spatial
resolution must be distinguished: they are related but different properties. Avoiding redundant
interpolation is an advantage of the computational arrangement studied; it does not establish that every
interface or plasma regime can be resolved at the same cost and accuracy.
A learning task based on a solver should distinguish the physical model, the discretisation and the
stopping rule. Refining a mesh while retaining the same boundary values tests a numerical choice;
changing the size of a truncated domain can test sensitivity to an artificial boundary; altering a material
law changes a physical assumption. These operations answer different questions even when all produce
changes in a plotted field.
Students can document one change at a time and state which claim their comparison supports. No
universal convergence rate should be assigned without the regularity and discretisation conditions
required by the method.

pág. 6046
Table 1 organises the formulations and approaches as possible objects of instruction. Its rows are not
competing categories at the same conceptual level: a particle model may use a finite-element field solve,
and a response surface may summarise outputs from that solve. The educational implication is to ask
learners to identify the role of each component before comparing accuracy or efficiency. This avoids
treating an algorithm name, a field equation and an optimisation strategy as interchangeable descriptions
of a computational model.
Table 1 Formulations, computational approaches and limits of comparison
Formulation or approach Represented quantity Limit relevant to teaching
Discrete Coulomb
superposition
Field from prescribed charges Induced charge requires additional
boundary constraints.
Poisson and Laplace
formulations
Potential subject to source, material
and boundary assumptions
Accurate potential values do not alone
establish accurate gradients.
Immersed FEM with PIC Field solution coupled to particle
evolution
Stability and resolution are different
properties.
Algebraic topological
method
Discrete relations in a boundary-value
problem
Exact identities do not imply zero
field error.
Response surface fitted to
FEM
Approximate design response Agreement with simulations is not
experimental or educational
validation.
Note. Synthesis of the formulations discussed above and the studies by Arun et al. (2023), Han et al. (2025), and Nadjem and
Rouagdia (2025). FEM = finite-element method; PIC = particle-in-cell. The teaching implications are interpretations of this
review, not measured learning outcomes.
Published observables and critical reading of computational evidence
In electrostatic separation, Nadjem and Rouagdia combine finite elements and response surface
methodology to explore electrode dimensions and inclination (Nadjem & Rouagdia, 2025). The study
reports an approximately 65% improvement in field parameters relative to comparison designs and close
agreement between simulations and statistical predictions. These values describe the selected design and
responses. They do not represent a 65% reduction in numerical error or, by themselves, demonstrate an
equivalent improvement in the purity or recovery of separated materials.
Tang et al. study multisource corona discharges through a three-dimensional fluid model and current
validation (Tang et al., 2026). Increasing the number of electrodes can increase total current while
reducing the mean current per needle through shielding. The comparison concerns the geometries and

pág. 6047
conditions investigated. A maximum normalised current error of 10% is not a 10% error in every local
field component: current integrates transport and ionisation processes and is a different observable.
Yu et al. present indirect electrostatic–fluid–structure coupling for a microelectromechanical comb
resonator (Yu et al., 2022). Their experimental comparison reports errors of 15.47% in amplitude and
12.48% in resonant frequency. These values must retain their reference variables and denominators
rather than becoming a generic measure of electromagnetic accuracy. Agreement in motion also depends
on damping, structural behaviour and residual thermal stress; adjusting only the field may leave a
dominant discrepancy unchanged.
Tsuchiya presents a device for quantitative visualisation of a two-dimensional distribution of
electrostatic vectors (Tsuchiya, 2021). It is a relevant measurement contribution, but its description is
not equivalent to complete real-time volumetric reconstruction of three components. Nor does it
establish that it is the only existing experimental method. This distinction matters because measurement
of one component, a projection or a planar map cannot automatically validate a three-dimensional field.
These cases can support an educational task in which students identify the numerator, denominator and
reference of an error statement before comparing results. A percentage referring to current cannot be
assigned to every field component, and an error in resonant motion does not isolate the error of a coupled
electrostatic calculation. Likewise, an improvement in a design response is not necessarily a reduction
in numerical error. Table 2 preserves these distinctions and can be used to discuss what evidence would
be needed for a stronger conclusion. Its instructional use is proposed here; the cited application studies
are not recast as educational experiments.

pág. 6048
Table 2 Reported observables and limits of inference in selected studies
Study Reported observable or result What cannot be inferred
Nadjem and Rouagdia
(2025)
Approximately 65% improvement in field
parameters relative to conventional
designs
An equivalent error reduction or
improvement in material recovery.
Tang et al. (2026) Maximum normalised current error of
10%
The same local error in every electric-
field component.
Yu et al. (2022) Amplitude error 15.47%; resonant-
frequency error 12.48%
Accuracy of the electrostatic solver
independently of the coupled model.
Tsuchiya (2021) Imaging of a two-dimensional distribution
of field vectors
Complete volumetric reconstruction
from the reported planar map.
Note. Source values are attributed to the studies in the first column. The final column states limits of inference applied in this
review. The quantities are not combined into a common performance indicator.
Teaching implications and evidence of explanation
The interface developed by Tan et al., the visualisations of Nascimento Neto and Siqueira, and
Villamizar’s didactic treatment make the relationship between potential, field and geometry explicit
(Nascimento Neto & Siqueira, 2025; Tan et al., 2024; Villamizar, 2025). Their representational
usefulness must be distinguished from metrological validation. Field-line density depends on graphical
seeding, normalised arrows can hide magnitude differences, and a planar section does not represent the
complete three-dimensional distribution. A scientific figure must identify the domain, units, scale,
charge positions and representation convention.
The synthesis suggests designing tasks around a relation that learners must explain, rather than around
a sequence of software commands. The intended product may be an annotated diagram, a comparison
of signed components, a boundary-condition argument or an explanation of an error metric. Each
product should expose the connection between a representation and a physical assumption. Table 3
translates the distinctions developed in the analysis into possible prompts and observable features of an
explanation. These features are proposed assessment criteria, not validated indicators or measurements
obtained from students.
A coherent classroom activity would make prediction, calculation and interpretation distinguishable.
Before running a model, students state a directional or qualitative prediction with its assumptions.
During calculation, they record the quantities varied and the conventions used to display the output.

pág. 6049
Afterwards, they compare the result with the prediction and explain any discrepancy using physical or
numerical reasons. This organisation is offered as a design option. The review does not prescribe a fixed
number of sessions or claim that the order is optimal for every population.
Assessment should preserve the difference between computational execution and physical explanation.
A functioning script can show that a procedure was implemented, but it does not by itself show that a
learner understands why the model is appropriate. Conversely, a minor programming error need not
invalidate a correct argument about superposition. A teacher can request both an interpretable output
and an explanation of one limiting case, then examine them separately. This makes it possible to locate
whether a discrepancy concerns vector reasoning, units, boundary assumptions, differentiation or the
implementation of an otherwise appropriate model.
Table 3 Proposed teaching tasks derived from the documentary synthesis
Analytical focus Possible task Evidence to examine
Scalar–vector distinction Compare the midpoint of equal-charge
and opposite-charge pairs
An explanation of scalar addition and
vector cancellation or reinforcement.
Potential and gradient Shift the reference potential and compare
field predictions
Recognition that a constant shift
preserves the field.
Boundary assumptions Contrast fixed potential with prescribed
surface charge
Identification of the specified
quantity and the induced unknown.
Spatial representation Label domain coordinates, plotted values
and units
Distinction between a potential
surface and a spatial domain.
Numerical verification Vary mesh resolution while retaining the
physical problem
Separation of discretisation
sensitivity from model changes.
Scope of evidence Interpret the different quantities reported
in Table 2
A conclusion limited to the
observable and comparison actually
reported.
Note. Author-proposed tasks derived from the preceding analysis. They have not been implemented or validated as a teaching
sequence or assessment instrument.
DISCUSSION
Educational contribution and limits of transfer
The contribution of the review is an organisation of explanatory relations for teaching electrostatics,
rather than a claim that visualisation alone improves conceptual understanding. The educational
antecedents address representational resources and teacher knowledge, while the disciplinary sources
impose constraints on what a representation may legitimately mean.

pág. 6050
Bringing them into dialogue makes it possible to distinguish an attractive display from an accountable
explanation. A task becomes relevant when it requires the learner to connect a physical assumption to
an operation and to the information preserved in the resulting representation.
This distinction also limits transfer from industrial or multiphysics applications. Their inclusion can
support scientific reading and analysis of evidence, but their reported performance does not validate a
teaching resource. A separator study, for example, may supply a question about design response; it does
not establish how students understand the electric field. Similarly, a measurement device can illustrate
the difference between a planar map and a volume, without proving that its use is practical or beneficial
in a particular classroom. Educational relevance requires an explicit instructional question, not merely
a sophisticated physical application.
For teacher education, the tables can serve as material for discussing why a representation is selected
and what a learner is expected to infer from it. The two TED studies provide examples of researching
teacher knowledge through educational evidence (Reyes Roncancio & Martínez, 2013; Viafara-Ortiz et
al., 2025). The present review contributes a disciplinary set of questions to that discussion, but it does
not characterise any teacher’s pedagogical content knowledge. Such a claim would require data from
actual teaching decisions and practice.
Physical consistency and the interpretation of learning evidence
Field recovery imposes an additional requirement beyond potential accuracy. Small spatial oscillations
in potential can produce appreciable changes in its gradient; a smooth plot or a small algebraic residual
therefore cannot establish local field accuracy. At interfaces, values must be evaluated from the
appropriate material side. For point charges, self-field is excluded when calculating force on the same
particle. If a singularity is regularised, the regularisation length and form must be stated because they
modify the effective model.
Error indicators must identify the variable, norm, domain and reference. Relative error becomes
misleading near a field null if the denominator is almost zero. Angular error is also undefined at such
points. Timing comparisons require equivalent tolerances and a description of mesh generation,
assembly, solution and post-processing costs. The examined literature contains useful results, but

pág. 6051
heterogeneous observables preclude calculating an aggregate improvement representative of every
methodological family.
These disciplinary checks can inform educational assessment without becoming a substitute for it. If
students are asked to explain a field null, their response should be interpreted with respect to the selected
configuration and representation. If they are asked about error, the reference and the observable must be
specified. A low score on a task with undefined symbols or an ambiguous colour convention may reflect
the task design rather than the targeted understanding. Any later investigation of the proposed activities
would need to document the instructional context and justify how the assessment captures the intended
reasoning.
The distinction between a conceptual survey and a locally designed modelling task is also relevant.
Maloney et al. (2001) provide a published approach to assessing conceptual knowledge in electricity
and magnetism. That work is an assessment antecedent, not evidence that the tasks in Table 3 are
validated. The latter target explanations involving particular modelling decisions and would require their
own examination. No reliability coefficient, achievement gain or causal effect can be assigned to them
from the present documentary synthesis.
Scope of the documentary conclusions
The review is selective and cannot establish which instructional approach is most effective across
courses or populations. Its initial working set favoured disciplinary and computational publications, and
the educational supplementation is purposive rather than exhaustive. The analysis therefore offers a
reasoned connection between physical constraints and possible teaching questions, not a complete state
of research on learning electrostatics. The difference is consequential: a topic insufficiently documented
in this selection cannot be described as absent from the literature.
The proposed tasks have not been applied to students. Their feasibility, difficulty and capacity to elicit
the intended reasoning remain empirical questions. Mathematical soundness does not establish
pedagogical effectiveness. The tables organise distinctions for instructional scrutiny; they are not
validated educational instruments.

pág. 6052
CONCLUSIONS
The educational reorientation of electrostatic modelling centres on the relationships learners must
explain among sources, potential, vector fields and computational representations. The documentary
analysis identifies distinctions that can guide those explanations: prescribed versus induced charge,
scalar values versus vector components, boundary assumptions versus software settings, and numerical
consistency versus physical or educational validation. Retaining these distinctions allows computational
work to be discussed as a modelling practice with explicit assumptions.
The proposed tasks connect analytical expressions, graphical conventions and the interpretation of
published results. They offer material for university physics teaching and teacher education, while
remaining proposals rather than demonstrated instructional outcomes. The review supports the selection
of explanatory questions and the delimitation of evidence; it does not establish learning gains, universal
superiority of three-dimensional modelling or an optimal teaching sequence. Its practical contribution
is a transparent basis for subsequent author-led design and empirical examination of teaching activities.
Author contribution
The four authors contributed to the conceptualisation of the study and the documentary analysis
underlying its conceptual framework. All authors participated in the scientific review and interpretation
of the literature, approved the final version of the manuscript for submission and accept responsibility
for its content.
Conflicts of interest and funding
The author declares no conflicts of interest. This work received no funding from public or private entities
or universities.
Declaration of generative AI assistance
ChatGPT was used for grammatical correction and linguistic editing of the manuscript. The authors are
responsible for the study design, the preparation of the article and the interpretation of the literature.
REFERENCES
Ainsworth, S. (2006). DeFT: A conceptual framework for considering learning with multiple
representations. Learning and Instruction, 16(3), 183–198.
https://doi.org/10.1016/j.learninstruc.2006.03.001

pág. 6053
Arun, A. D., Thirunavukkarasu, S., Chandra, S., Rajiniganth, M. P., Malathi, N., & Sivaramakrishna,
M. (2023). Investigating algebraic topological method for solving 3D Laplace equation in
electrostatic boundary value problems. Computer Physics Communications, 289, 108759.
https://doi.org/10.1016/j.cpc.2023.108759
Ballas, R. G. (2025). Energy density in the electrostatic field. In Piezoelectric Bending Transducers
(pp. 121–167). Springer. https://doi.org/10.1007/978-3-662-70689-3_6
Becerra Becerra, A., Hernandez Naranjo, A., & Vargas Vera, C. (2026). Finite-point electrostatic
representation of spherical and non-spherical charged particles. Contributions to Plasma
Physics, e70171. https://doi.org/10.1002/ctpp.70171
Carrillo Avila, A. R. (2025). La permitividad en electrostática a partir del análisis dimensional. Revista
de Enseñanza de la Física, 37(1), 125–134. https://doi.org/10.55767/2451.6007.v37.n1.48980
Cho, M. (2022). A novel efficient numerical solution of Laplace equation with mixed boundary
conditions. International Journal of Computer Mathematics, 99(6), 1272–1280.
https://doi.org/10.1080/00207160.2021.1967939
Han, Y., Xia, G., Kang, H., Lu, C., Chen, C., & Saetchnikov, V. (2025). Three-dimensional direct-
implicit particle-in-cell model using trilinear anisotropic immersed-finite-element for plasma
propulsion. AIAA Journal, 63(1), 323–335. https://doi.org/10.2514/1.j064305
Liao, Z., Xu, J., Wan, B., Gan, Z., Wang, Y., Li, N., Liu, Y., & Long, T. (2023). Finite element analysis
of total electric field distribution on the roof of flat-roofed buildings near HVDC lines.
Electrical Engineering, 105(2), 1153–1162. https://doi.org/10.1007/s00202-022-01722-1
Lima, F. M. S. (2025). On the electric field at the surface of a uniformly-charged cylindrical shell.
Revista Brasileira de Ensino de Física, 47, e20240312. https://doi.org/10.1590/1806-9126-rbef-
2024-0312
Maloney, D. P., O’Kuma, T. L., Hieggelke, C. J., & Van Heuvelen, A. (2001). Surveying students’
conceptual knowledge of electricity and magnetism. American Journal of Physics, 69(S1), S12–
S23. https://doi.org/10.1119/1.1371296

pág. 6054
Montilla López, Y., Suárez Carreño, F., & Linares Rodríguez, B. (2025). Visualización del campo
electrostático con Excel y Maple para la enseñanza en ingeniería. Revista Brasileira de Ensino
de Física, 47, e20240362. https://doi.org/10.1590/1806-9126-rbef-2024-0362
Nadjem, A., & Rouagdia, K. (2025). A combined RSM-FEM analysis of electric field distribution in a
novel design of an inclined-plane electrostatic separator. Applications of Modelling and
Simulation, 9, 199–212. https://arqiipubl.com/ojs/index.php/AMS_Journal/article/view/876
Nascimento Neto, O. J., & Siqueira, M. R. S. (2025). Explorando a equação de Laplace visualmente:
uma ferramenta para o ensino de eletromagnetismo. Revista Brasileira de Ensino de Física, 47,
e20240404. https://doi.org/10.1590/1806-9126-rbef-2024-0404
Polito, A. M. M. (2025). Localizability of charge and current distributions in the regime of
electromagnetostatics and the mathematical proofs of Coulomb and Biot-Savart laws. Revista
Brasileira de Ensino de Física, 47, e20250204. https://doi.org/10.1590/1806-9126-rbef-2025-
0204
Reyes Roncancio, J. D., & Martínez, C. A. (2013). Conocimiento didáctico del contenido en la
enseñanza del campo eléctrico. Tecné, Episteme y Didaxis: TED, (33).
https://doi.org/10.17227/01213814.33ted36.60
Salazar, R., Bayona-Roa, C., & Jaramillo, D. (2022). Electric vector potential and the Biot-Savart like
law in electrostatics. Revista de la Academia Colombiana de Ciencias Exactas, Físicas y
Naturales, 46(181), 939–946. https://doi.org/10.18257/raccefyn.1671
Tan, X. Y., Chan, K. S., & Lim, S. Y. (2024). Development of a MATLAB‐based graphical user
interface to illustrate Coulomb’s law and Gauss’s law. Computer Applications in Engineering
Education, 32(5), e22759. https://doi.org/10.1002/cae.22759
Tang, Z., Li, C., Li, D., Ma, C., Wei, E., Liang, T., Xiao, M., Zhou, Y., Zhao, X., & Du, Y. (2026).
Three-dimensional fluid dynamics simulation of cross-scale multi-needle corona discharges.
Plasma Processes and Polymers, 23(7), e70216. https://doi.org/10.1002/ppap.70216
Tsuchiya, M. (2021). Imaging of electrostatic field vector distribution. Review of Scientific Instruments,
92(5), 053704. https://doi.org/10.1063/5.0034052
pág. 6055
Viafara-Ortiz, R., Cardona-Prato, A., & Montaño-Duarte, A. (2025). Exploración del conocimiento
pedagógico del contenido en un docente novel de física. Tecné, Episteme y Didaxis: TED, (58),
205–221. https://doi.org/10.17227/ted.num58-22073
Villamizar, D. V. (2025). A didactical approach to teaching electric field and electric potential concepts.
Revista Brasileira de Ensino de Física, 47, e20250334. https://doi.org/10.1590/1806-9126-rbef-
2025-0334
Wang, Q., Liu, L., Liu, D., & Wang, H. (2022). Research on two-dimensional electrostatic particle
simulation method based on superLU. In 2022 23rd International Vacuum Electronics
Conference (IVEC) (pp. 468–469). IEEE. https://doi.org/10.1109/ivec53421.2022.10292209
Yu, Z., Chen, S., Mou, Y., & Hu, F. (2022). Electrostatic-fluid-structure 3D numerical simulation of a
MEMS electrostatic comb resonator. Sensors, 22(3), 1056. https://doi.org/10.3390/s22031056