Tachyon condensation can stabilise bosonic string theory inside curved spacetime
A newly calculated positive quartic term suggests that rolling tachyons in bosonic string theory settle into a stable Anti-de Sitter geometry rather than triggering runaway collapse.

A physical system balanced on a peak will inevitably slide downward. In quantum field theory, an empty region of space behaves the same way when the fields that define it carry negative squared masses. Such fields, known as tachyons, signal that the assumed vacuum configuration is not a true resting point but a temporary summit. Instead of remaining empty, spacetime experiences runaway instability as the field rolls downhill, shedding energy and fundamentally altering the geometry of the surrounding universe.
For more than half a century, this instability cast bosonic string theory as an incomplete toy model.1 The formulation describes fundamental particles as tiny vibrating loops of string in twenty-six dimensions, yet its lowest-energy state is a tachyon with negative mass squared.2 Because free strings in flat Minkowski spacetime possess this imaginary-mass excitation, physicists widely viewed the flat vacuum of bosonic string theory as fatally unstable. Most researchers turned their attention to superstring theory, where supersymmetry eliminates tachyonic states entirely and preserves a stable vacuum in flat dimensions.13
A preprint posted to the arXiv repository on September 7, 2026, by Mohammad R. Garousi, a theoretical physicist in the Department of Physics at Ferdowsi University of Mashhad, explores an alternative fate for the non-supersymmetric bosonic string. By calculating higher-order interactions between tachyons and massless particles, Garousi demonstrated that the closed string tachyon does not roll into endless destruction. The potential energy curve turns upward, providing a stable minimum that reconstructs spacetime into a negatively curved Anti-de Sitter geometry.1
Why does the presence of a tachyon destabilise spacetime?
Spacetime becomes unstable in the presence of a tachyon because the quadratic term in the field's potential energy carries a negative sign, causing any small fluctuation to grow exponentially over time. To understand why this happens, consider the causal chain that governs physical fields. First, a field across spacetime requires energy to exist, and its mass term sets the baseline energy of uniform configurations. Second, when the mass squared is positive, displacing the field away from zero costs energy, pulling it naturally back toward equilibrium. Third, when the mass squared is negative, displacing the field lowers its potential energy, transforming zero into an unstable maximum. Finally, the field accelerates down the slope, releasing energy that must either settle into a new, lower valley or tear the background geometry apart.
In bosonic string theory, the mass squared of the closed string tachyon is fixed at negative four divided by the string parameter alpha prime, which sets the characteristic length scale of string interactions. Setting the convention of alpha prime to two gives a tachyon mass squared of negative two.1 Around a flat twenty-six-dimensional Minkowski background, this negative value means that the empty vacuum cannot persist. As the tachyon field rolls away from zero, it interacts with the massless excitations of the string: the spacetime metric describing gravity, the scalar dilaton determining the strength of string interactions, and the antisymmetric Kalb-Ramond tensor field.2
Previous studies in string theory established that open string tachyons, which describe the vibrations of strings attached to boundary surfaces called D-branes, condense to a stable ground state that cancels the brane tension.14 The fate of closed string tachyons, which describe vibrations of unattached string loops throughout the entire bulk of spacetime, proved far harder to determine.2 Investigating whether the closed string tachyon halts its descent requires examining the higher-order terms of its potential, beginning with the four-field interaction term.1 If the coefficient of this quartic term is positive, the energy curve turns back upward, creating a basin where the field can safely come to rest.5
How does T-duality eliminate ambiguities in string field equations?
T-duality resolves ambiguities in string field equations by requiring that physical interactions remain identical when a spatial dimension of radius R is exchanged for one of radius alpha prime divided by R.12 In string theory, the classical scattering amplitude, known as the S-matrix, describes how external particle states scatter off one another.1 However, extracting an unambiguous effective action from scattering amplitudes is notoriously difficult.15 When particles satisfy their on-shell equations of motion, field redefinitions can shift terms between different derivative structures.1 For instance, a term coupling spacetime curvature directly to two derivatives of the tachyon can be rewritten as a term coupling curvature to the square of the tachyon without derivatives.1

Garousi resolved this ambiguity by establishing a specific field-redefinition scheme.5 Through higher-derivative shifts of the tachyon field, all couplings between the tachyon and massless fields can be systematically reorganized so that the tachyon appears without derivatives exclusively within the tachyon potential.1 In this chosen scheme, every other interaction term in the action involves only covariant derivatives of the tachyon, metric, dilaton, and antisymmetric tensor.5 This convention prevents higher-derivative corrections from blurring into the potential itself, providing a clean separation between the shape of the potential and derivative-dependent kinematic interactions.1
Asked whether altering the choice of field variables could shift this potential or change its sign, Mohammad said: "A different choice of field variables may not resolve the on-shell ambiguity that the Laplacian of the tachyon can be replaced by a non-derivative tachyon term. I think that for the field variables in which the non-derivative tachyon appears only in the tachyon potential, the sign of the quartic term in the tachyon potential would not be changed, because this term is fixed by comparing with the sphere-level string theory S-matrix element of four tachyons."
To construct the effective action at four-derivative order, corresponding to order alpha prime, Garousi identified all possible scalar contractions of the Riemann curvature tensor, the Kalb-Ramond field strength, the dilaton gradient, the tachyon gradient, and their covariant derivatives.1 Accounting for Bianchi identities, total derivative terms, and field redefinitions of the metric, dilaton, and Kalb-Ramond field, an initial set of 56 potential couplings collapsed into a minimal basis of 14 independent terms.1 In accordance with an earlier proposal by Garousi, the basis excluded all couplings containing an odd number of tachyon fields.14 Although string scattering amplitudes with an odd number of external tachyons are nonzero, consistency with general covariance and the pole structure of the S-matrix requires that the spacetime effective action contain only even powers of the closed string tachyon.14
Explaining how non-vanishing odd amplitudes are reconciled with an action containing only even powers, Mohammad said: "Couplings with an odd number of tachyons appear in the S-matrix elements of an even number of tachyons as poles. To reproduce a generally covariant effective field theory, these poles must be expanded to produce higher-momentum terms, which eventually generate higher-derivative couplings of an even number of tachyons in the effective action. The same situation occurs for couplings of massive fields. The S-matrix elements involving massive fields are nonzero. However, in the S-matrix elements of massless fields, the massive fields appear as poles, and one expands these poles to produce higher-momentum terms that are ultimately reproduced by higher-derivative couplings of massless fields in the effective action."
Garousi then imposed the symmetry constraint of T-duality on this 14-term basis.15 By compactifying one spatial direction on a circle and applying the Buscher transformation rules supplemented by higher-derivative corrections, the reduced twenty-five-dimensional action must remain invariant.12 Applying this constraint across a flat base space established rigorous relations among the coupling constants.1 T-duality forced eight of the 14 coefficients to zero and expressed the remaining six non-zero couplings in terms of two parameters. One parameter was determined by matching the couplings to the established zero-tachyon effective action in the Metsaev-Tseytlin scheme.5 The single remaining parameter, multiplying the four-tachyon derivative interaction, was fixed by comparing the action against the sphere-level string scattering amplitude of four tachyon vertex operators.15
What happens when the tachyon reaches the bottom of its potential?
When the tachyon rolls to the bottom of its newly calculated potential, it settles at a finite field value that endows spacetime with a constant negative vacuum energy density.1 The comparison between the field theory action and the low-energy expansion of the four-tachyon sphere scattering amplitude did more than fix the final derivative coupling. The zero-momentum pieces of the four-tachyon amplitude simultaneously determined the quartic coefficient in the tachyon potential.1 Garousi found that this coefficient is strictly positive, verifying that the potential does not plunge indefinitely into negative infinity.15
Taking the tachyon mass term of negative one-half T squared alongside the newly derived positive quartic term reveals a striking mathematical pattern. Garousi noted that these first two terms match the Taylor series expansion of a smooth cosine potential proportional to two divided by alpha prime multiplied by the quantity negative one plus cosine of T.1 In this periodic function, the unstable flat vacuum sits at T equals zero, where the potential energy is zero and the curvature of the potential is negative.1 As the tachyon rolls away from zero, it descends toward a stable local minimum located at T equals pi.1

However, the periodic cosine profile remains an unproven hypothesis. Asked about the physical motivation behind this periodic curve, Mohammad said: "The form \(\cos T\) is merely a speculation, based on the fact that this function reproduces the correct \(T^2\) and \(T^4\) couplings."
At this conjectured minimum, the tachyon field attains a non-zero vacuum expectation value.1 Evaluating the cosine potential at T equals pi yields a negative energy density of negative four divided by alpha prime.1 In general relativity, a uniform negative potential energy acts as a negative cosmological constant, denoted by Lambda equals negative four divided by alpha prime.1 The backreaction of this negative energy curves the twenty-six-dimensional spacetime into an Anti-de Sitter geometry.1 Condensation therefore transforms an unstable flat universe into a stable, maximally symmetric curved spacetime.15
The findings provide a different lens on the long-standing puzzle of vacuum stability in bosonic string theory. In a related context, Matilda Delgado at the Instituto de Fisica Teorica in Madrid showed that instabilities like the bubble of nothing can be understood through closed string tachyon condensation terminating spacetime geometry.1 Asked where the boundary lies between geometric termination and stabilization into Anti-de Sitter space, Mohammad provided no clear insight into the question. Similarly, when asked what specific physical conditions or energy scales might prevent the transition to curved space, Mohammad had no clear answer to give.
What limits the validity of the curved ground state?
The calculations represent a classical field theory derivation at four-derivative order rather than an exact non-perturbative proof of string dynamics.1 The analysis operates at the sphere level of string perturbation theory, meaning it ignores quantum string loops and multi-loop corrections to the effective potential.1 Furthermore, the cosine potential is an extrapolation based on matching the quadratic mass term and the quartic interaction coefficient, leaving open the possibility that six-point and higher-order scattering amplitudes introduce higher-order corrections. The dimensional reduction used to impose T-duality also assumed a flat lower-dimensional base space.1
Quantum effects could alter this classical picture significantly. Asked what specific higher-genus string loop calculation could overturn the result, Mohammad said: "The one-loop dilaton tadpole is nonzero in bosonic string theory. Including this effect may modify the classical result." In closed bosonic string theory, the dilaton tadpole generates infrared divergences at loop level that reflect the backreaction of quantum vacuum fluctuations on spacetime geometry.7
Testing whether the cosine potential survives higher-derivative scrutiny forms the immediate next step. Calculating the six-tachyon scattering amplitude would determine the coefficient of the sixth-order tachyon term, confirming or modifying the trigonometric form. If the Anti-de Sitter minimum remains stable under quantum corrections, bosonic string theory could offer a tractable laboratory for studying quantum gravity, tachyon condensation, and gauge-gravity duality without relying on supersymmetry.1
This piece was prepared from the arXiv preprint and public records together with answers from Mohammad to six questions from the Primary news team, completed September 2026.1
The record
57 sentences trace to 6 sources and 1 contributor.
- 1 Tachyon-massless couplings at order $\alpha'$ in bosonic string theory Preprint · may not have been peer reviewed See the source
- 2 The Bubble of Nothing under T-duality Preprint · may not have been peer reviewed See the source
- 3 Preface Preprint · may not have been peer reviewed See the source
- 4 1The basic idea Preprint · may not have been peer reviewed See the source
- 5 Contents Preprint · may not have been peer reviewed See the source
- 6 Contribution — Mohammad R. Garousi 8 statements added to this article
- 7 Contents Preprint · may not have been peer reviewed See the source
Other consulted pages1 page
ar5iv.labs.arxiv.org1 page
arxiv.org5 pages
Article history
-
Published 17 Nov 2026, 19:42Assembled by the Primary desk from 6 sources · 1 contributor · 57 cited sentences
-
Mohammad R. Garousi, a theoretical physicist in the Department of Physics at Ferdowsi University of Mashhad
- Go to the sentence: Through higher-derivative shifts of the tachyon field, all couplings between the tachyon and massless fields can be systematically reorganized so that the tachyon appears without derivatives exclusively within the tachyon potential.
- Go to the sentence: Asked whether altering the choice of field variables could shift this potential or change its sign, Mohammad said: "A different choice of field variables may not resolve the on-shell ambiguity that the Laplacian of the tachyon can be replaced by a non-derivative tachyon term. I think that for the field variables in which the non-derivative tachyon appears only in the tachyon potential, the sign of the quartic term in the tachyon potential would not be changed, because this term is fixed by comparing with the sphere-level string theory S-matrix element of four tachyons."
- Go to the sentence: Explaining how non-vanishing odd amplitudes are reconciled with an action containing only even powers, Mohammad said: "Couplings with an odd number of tachyons appear in the S-matrix elements of an even number of tachyons as poles. To reproduce a generally covariant effective field theory, these poles must be expanded to produce higher-momentum terms, which eventually generate higher-derivative couplings of an even number of tachyons in the effective action. The same situation occurs for couplings of massive fields. The S-matrix elements involving massive fields are nonzero. However, in the S-matrix elements of massless fields, the massive fields appear as poles, and one expands these poles to produce higher-momentum terms that are ultimately reproduced by higher-derivative couplings of massless fields in the effective action."
- Go to the sentence: Asked about the physical motivation behind this periodic curve, Mohammad said: "The form \(\cos T\) is merely a speculation, based on the fact that this function reproduces the correct \(T^2\) and \(T^4\) couplings."
- Go to the sentence: Asked where the boundary lies between geometric termination and stabilization into Anti-de Sitter space, Mohammad provided no clear insight into the question.
- Go to the sentence: Similarly, when asked what specific physical conditions or energy scales might prevent the transition to curved space, Mohammad had no clear answer to give.
- Go to the sentence: Furthermore, the cosine potential is an extrapolation based on matching the quadratic mass term and the quartic interaction coefficient, leaving open the possibility that six-point and higher-order scattering amplitudes introduce higher-order corrections.
- Go to the sentence: Asked what specific higher-genus string loop calculation could overturn the result, Mohammad said: "The one-loop dilaton tadpole is nonzero in bosonic string theory. Including this effect may modify the classical result."

