WorksPhysics and Astronomy
Polymer adsorption on vicinal surfaces
Abstract
While polymer adsorption is theoretically well established on flat surfaces — and somewhat understood on rough surfaces — the common and practically relevant case of nearly flat stepped surfaces remains poorly studied. Here, we address this gap via extensive simulations of flexible polymer chains grafted at the step edge of vicinal surfaces. For wide terraces, the polymer-step interaction yields a stable step-attached (SA) phase, alongside the desorbed and terrace-adsorbed (TA) phases, all of them separated by continuous transition lines that meet at a multicritical point previously unanticipated in polymer thermodynamics. This scenario is observed for terrace widths , being the radius of gyration of TA polymers. Simpler phase diagrams, absent this multicriticality, are found for , where a simple crossover exists between the SA and TA phases. The presence of kinks at the steps does not alter these behaviors. Besides their theoretical relevance, these findings unveil an interesting route to produce polymer nanowires and control polymer conformation via step attachment-detachment on designed stepped surfaces.
Introduction
Polymer adsorption at surfaces is key to a vast number of industrial and biological processes, including adhesion [1], lubrication [2], stabilization of colloidal suspensions [3], fabrication of biosensors [4], drug delivery systems [5], and many others [6]. This has motivated a large body of literature aimed at understanding the underlying physics of polymer adsorption. It is well known that when a polymer is exposed to an attractive surface, the competition between the energetic gain of attachment and the entropic loss of surface proximity drives the system through transitions from desorbed to adsorbed phases, depending on the polymer-substrate interaction strength () and the temperature () [7, 8, 9]. The great majority of these studies consider adsorption on perfectly flat and homogeneous surfaces [7, 8, 9]; yet, even this ideal situation remains a topic of recent debate regarding the critical properties at the ordinary [10, 11, 12, 13] and special adsorption transitions [13, 14].
Surface inhomogeneities may also play a pivotal role in the adsorption process. This has been demonstrated, for instance, in several works theoretically analyzing polymer adsorption on perfectly flat but chemically inhomogeneous surfaces — formed by different functional groups arranged in checkerboard [15], stripe [16], and other regular patterns [17, 18, 19], as well as in random configurations [20] — which are relevant for molecular recognition [21].
Though similarly important, the effects of physical inhomogeneities (arising from inexorable surface imperfections) on adsorption are not as well explored. Indeed, most existing theoretical works focus on locally smooth but globally rough planar surfaces featuring, for example, sinusoidal [22, 23] or checkerboard [24] corrugations, or square-well holes [25]. The same is true for polymer adsorption on non-planar geometries, such as the smooth surfaces of spheres [23, 26], cylinders and rods [7, 27], mesoscopic particles [28], and repulsive wedges with an attractive edge [29]. Adsorption on planar substrates with disordered and fractal shapes has also been analyzed in a few studies [30].
At a microscopic level, solid surfaces are rarely atomically flat, nor do they always exhibit a large roughness. Often, the closest approximation to a flat system is a low-index facet featuring broad flat terraces separated by long steps — so-called vicinal surfaces [31]. Arrangements of steps and terraces are indeed ubiquitous in crystalline surfaces [31, 32]. While the role of such steps in polymerization has been discussed in the literature [33], their effect on polymer adsorption has remained largely unappreciated, except for very short chains [34]. We remark that when adatoms (atoms or small molecules) adsorb onto such surfaces and subsequently diffuse, the steps are critical to their aggregation process; among other reasons [35], this is because adatoms interact with the step in addition to the adatom-terrace interaction. Naturally, this step-induced effect should be negligible for long polymers if their persistence length, , is much larger than the step height, . However, in many realistic scenarios, one may find , even for single atomic steps, where – nm for metallic and semiconductor surfaces [37]. Relevant examples of polymers with such a small include polyethylene glycol (PEG, nm) [38], unfolded or intrinsically disordered proteins ( nm) [39], polydimethylsiloxane (PDMS, – nm) [40], and single-stranded DNA ( nm) [41]. Moreover, can be much larger than this due to step bunching or on microfabricated surfaces. Therefore, steps can dictate the macromolecular conformation at the surface for a broad class of polymers, offering an avenue to control their properties via the substrate.
In this Letter, we tackle this fundamental issue through a detailed numerical study of the adsorption of flexible linear homopolymer chains onto vicinal surfaces under dilute, good-solvent conditions. Analogous to adsorbed adatoms, the polymer monomers at the surface interact attractively with both the terraces and the steps. We investigate vicinal surfaces with straight and kinked steps, for several terrace widths. Importantly, in addition to the well-known desorbed and terrace-adsorbed phases, we identify a stable step-attached phase and a previously unobserved desorbed-to-step-attached continuous phase transition, with unexpectedly large critical exponents. The phase diagrams of the system are fully characterized, revealing distinct behaviors for short versus wide terraces, and exposing a novel multicritical behavior for polymer adsorption in the latter case.
Model and Methods. As is standard in the literature [42, 43], we model swollen polymer chains in solution as self-avoiding walks (SAWs) comprising monomers on a simple cubic lattice. This lattice has a spacing and is effectively infinite in all directions, but it is bounded from below by an impenetrable surface composed of terraces (forming a square lattice of adsorption sites). These terraces have a length and a width , and are separated by parallel steps. Each step has a height and features kinks separated by a distance [see Fig. 1(a)]. The physical scale of interest here is such that the polymer’s persistence length and the step height are both comparable to the lattice spacing, meaning that . One end of the polymer is grafted at a step edge, in the center of the step.
As illustrated in Figure 1(b), each monomer residing on the terrace is assigned a statistical weight while those located at the step edge also carry a weight , where is the Boltzmann constant and the energies and are both negative. Monomers positioned at kink sites make two distinct contacts with the

Figure 1. (a) Schematic representation of a vicinal surface with terrace width and steps of height , length and distance between kinks . (b) Lateral view of a polymer near a step. Each monomer (black circle) contact with a terrace site is indicated by a red arrow, while the green arrow represent a contact with a step. (c) Definition of the coordinate system.
step, thus contributing a weight of . Let denote the total number of SAWs of length having exactly step contacts and terrace contacts. The partition function of the system can then be written as
and the expected value of any physical observable is given by
where the sum is evaluated over all possible SAW configurations . We estimated these thermodynamic averages by performing extensive numerical simulations using the stochastic growth algorithms PERM [44] and flat-PERM [45]. In both approaches, a chain is grown from an initial monomer (located at the origin of the coordinate system), with each subsequent monomer being added to a randomly chosen, unoccupied neighboring site of the previous one. FlatPERM and PERM both employ pruning and enrichment strategies to reduce attrition and control the weight dispersion during the SAW growth process. The flatPERM algorithm provides direct estimates of , which enables us to thoroughly explore the full parameter space for short-length SAWs ( here). Conversely, we utilize PERM to perform a more focused analysis at fixed values of and , which allows us to study much longer chains ( here). For each parameter set, walks were generated.
Wide terraces with straight steps. We begin by discussing the results for straight steps separating broad terraces, ensuring that the polymer interacts with only a single step. This regime corresponds to and , where is the average radius of gyration, is the Flory exponent for 2D SAWs [46], and the amplitude is for a square lattice [47]. As illustrated in Figure 2, three distinct stable phases are identified within the phase diagram:

Figure 2. Examples of polymer configurations for in the (a) desorbed (), (b) terrace-adsorbed (), and (c) step-attached () phases. Red, gray, and white spheres represent desorbed, terrace-adsorbed, and step-attached monomers, respectively.
For small values of and , the system is characterized by small averages for both and . This corresponds to the desorbed phase, where the polymer chain spans the 3D space [Fig. 2(a)]. ii) For large and small , is high, but the chain exhibits very few step contacts. Hence, the polymer adopts quasi-2D conformations in this terrace-adsorbed (TA) phase [see Figure 2(b)].
iii) For large , is large, while the difference remains small, meaning that most adsorbed monomers reside at the step edge. As shown in Figure 2(c), the chains are quasi-1D in this step-attached (SA) phase.
This three-phase scenario is confirmed both by fluctuation maps for short chains—which correspond to density plots of the largest eigenvalue of the covariance matrix for and , and yield approximations for the phase diagrams of the system; see Figure S6 of the Supplemental Material (SM) [50]—and by various other observables obtained from PERM simulations for long chains. For instance, a reliable way to determine the adsorption transition lines is via the scaling of the perpendicular and parallel components of the mean squared end-to-end distance, . In our specific case, however, we must analyze the directional components:

Figure S6. Results for vicinal surfaces with straight steps and short terraces. Fluctuation maps in the – plane for a polymer length and terrace widths of (a) , (b) , (c) , (d) , and (e) .
where are the coordinates of the free end of the chain, as defined in Figure 1(c). In the asymptotic limit, we expect to find: in the desorbed phase; and in the TA phase; and with in the SA phase.
Figure 3(a) shows that when is small, all exponents take values close to that for SAWs on 3D lattices ( [48]) at small , but subsequently shift to and for large .
Figure 3. Effective Flory exponents versus () for fixed values of () on a vicinal surface with wide terraces and straight steps. The chain lengths are (a)–(c) and (d) .
This indicates that a phase transition from the desorbed to the TA phase occurs at , corresponding to the crossing point of the exponent curves. The exact same behavior is observed for with an identical transition point within our precision, establishing that a desorbed-to-TA transition line exists in the phase diagram at [see Figure 4(a)]. Notably, this matches the critical point for adsorption on a perfectly flat 2D surface [49], which remains unaltered by the presence of the step, provided that . Further evidence supporting this scenario, derived from other quantities, is presented in the SM, where we also demonstrate that the critical exponents and (see Sec. III and Figure S1 of the SM) calculated along this line are entirely consistent with those for flat substrates, i.e., [11, 49].
Figure 4. Phase diagrams in the – plane for polymer chains tethered on vicinal surfaces with wide terraces separated by (a) straight () and (b) kinked () steps.
Figure S1. Results for vicinal surfaces with broad terraces and straight steps. (a) Effective exponents (dashed) and (solid lines), and (b) fluctuations as functions of , for . Data for different chain lengths are shown in both panels, as indicated by the legend in (a). The inset in (b) presents the maxima of against .
For , we find when is small. For large , however, Figure 3(b) demonstrates that while , as expected for the SA phase. Therefore, increasing at fixed values of drives the system through a desorbed-to-SA transition. An analogous behavior occurs if one holds fixed and increases . The resulting critical line is depicted in the phase diagram of Figure 4(a), having a maximum value () at , above which only the SA phase remains stable. This novel transition is confirmed via other thermodynamic quantities in Sec. III-B of the SM, where the critical exponents are estimated. Although this large value might hint at a first-order transition, no such evidence is observed in the histograms for the number of step contacts [Figure S2, SM], strongly suggesting that the desorbed-to-SA transition is continuous.
Figure S2. Results for vicinal surfaces with broad terraces and straight steps. (a) Fluctuations versus , for the indicated lengths . (b) Distribution of monomer-step contacts at different values of near the critical point, for . The results in panels (a) and (b) are for , and similar behaviors are found for other . Panels (c) and (d) show the extrapolations of the pseudocritical exponents , , , and obtained along the desorbed to step-attached transition line for fixed parameters indicated in the legend of (c).
Fixing and increasing , as shown in Figure 3(c), yields and for small , and with for large . Hence, the system undergoes a TA-to-SA transition. From the intersections of these Flory exponents, we map the corresponding transition line starting at a multicritical point (where it meets the other two critical lines mentioned above) and extending toward . In this asymptotic limit, the polymer becomes fully surface-adsorbed, recovering the well-studied case of 2D polymer adsorption onto a 1D boundary line; this transition is continuous, occurring at with critical exponents [51]. These same exponents, as well as consistent estimates for the critical points, are observed here for all . For smaller , however, the exponents measured conventionally begin to deviate from these expected values, and these deviations grow more pronounced as the multicritical point is approached [Fig. S3, SM]. This behavior occurs because a considerable fraction of the monomers desorbs when , a fraction that systematically increases with decreasing . Indeed, by substituting the chain length with the actual number of adsorbed monomers within the scaling relations describing the adsorption, we recover values of and that fall significantly closer to [Fig. S3, SM].
Such pronounced finite-size effects prevent us from accurately locating the multicritical point and its associated critical exponents. Nonetheless, assuming it resides at , as indicated by the results above, the crossing of versus along this line yields an estimated value of [see Fig. 3(d)].
Wide terraces with kinked steps. To investigate the effect of kinks on polymer adsorption, we consider vicinal surfaces with and . We start focusing on the limiting situation . The coordinate system is defined here analogously to Figure 1(c), with the - and -directions being perpendicular and parallel to the step, respectively, and perpendicular to the terrace. The Flory exponents for kinked steps [Fig. S4, SM] exhibit variations very similar to those displayed in Figure 3 for the kink-free case. Results for other observables, including the critical exponents, are also analogous to those discussed above for straight steps. Overall, the phase diagrams for polymer adsorption on surfaces with straight and kinked steps with are almost identical [see Fig. 4(b)]. A slight upward shift is observed only in the TA-to-SA transition line for the kinked case, as well as in the multicritical point, which is located at . The same picture is obtained for , but with intermediate TA-to-SA transition lines and , demonstrating that kinks exert a minor effect on the adsorption process.
Short terraces with straight steps. Let us thus return to straight steps, considering vicinal surfaces with small terrace widths such that the polymer can cross and attach to different steps. Once again, we consider the polymer grafted at a step edge, although similar results are found for short terraces if we move its initial position to the middle of a terrace.
When , which corresponds to step bunches forming a -sloped staircase, every monomer at the surface simultaneously makes one contact with a terrace and one with a step. Consequently, the SA and TA phases
become indistinguishable, yielding a single desorbed-to-adsorbed transition, as illustrated by the fluctuation map for [Fig. S5, SM]. This transition line occurs at , along which the critical exponents are for [Fig. S5, SM]. Remarkably, even though the polymer temporarily loses contact with the surface when crossing a step, its scaling behavior remains identical to that of perfectly flat surfaces.
Fluctuation maps for widths and [Fig. S6, SM] indicate that the results for are highly similar to those for , though the boundary of the desorbed phase becomes asymmetric with respect to the variables and . For , this boundary is nearly identical to those obtained for larger widths, and a bright stripe separating the regions corresponding to the TA and SA phases emerges in the map. However, the fluctuations remain quite mild along this stripe, indicating a crossover between these two phases, although these fluctuations get stronger as increases.
It is important to note that keeping fixed while increasing causes the polymer to visit fewer steps, a situation analogous to keeping fixed while decreasing . Therefore, to unveil the asymptotic adsorption behavior for short terraces, the finite-size analysis must be performed by keeping the number of visited steps fixed. Equivalently, we discuss this behavior in terms of the ratio . For instance, Figure 5(a) presents the variation of the Flory exponents with for , , and (matching the case for and ). Similarly to Figure 3(c), and exhibit clear signatures of the TA and SA phases at small and large , respectively; but there is no crossing of exponents’ curves, consistently with a crossover. (Note that is non-zero in the TA phase here because the grafting origin is fixed on a single terrace.) Figure 5(b) compares the same exponent curves for at lengths () and (). While no crossing is observed in the latter case, a clear intersection is present for . Results for other observables [shown in Figure S7, SM] confirm that only a TA-to-SA crossover exists when . For example, we find that deep inside both the TA and SA phases for such small ratios, implying that the adsorbed chains always maintain a macroscopic number of step contacts; increasing merely causes them to localize on fewer steps without altering their fundamental scaling behavior. Moreover, similar to the Flory exponents discussed above, we find that curves of versus for exhibit intersection points characteristic of a critical transition for (corresponding to ). However, the curves for and (for which and , respectively) do not cross (either each other or the other curves), confirming that the criticality changes to a crossover when [Figure S7, SM]. On the other hand, by increasing while keeping this ratio fixed at , for example, the signature of criticality appears for all simulated lengths [Figure S7, SM]. This strongly suggests that the spatial constraint separating the crossover and critical behaviors is at . Additional evidence comes from the exponents measured deep inside the TA phase for different ratios [Figure S7, SM]. Their effective values stop depending on this ratio and approach zero when , demonstrating that this is the condition for the number of monomer-step contacts becoming negligible in the TA phase, allowing for the onset of a critical TA-to-SA transition.
Figure 5. Effective Flory exponents versus , for . Data for and are shown in (a), while panel (b) presents results for a wider terrace and different chain lengths. The symbols in (b) follow the legend in (a).
Figure S7. Results for vicinal surfaces with straight steps and short terraces. Panels (a)–(c) plot the effective exponent versus for: (a) and ; (b) and ; and (c) utilizing varying widths specifically chosen to maintain . Panel (d) demonstrates the variation of against the ratio , evaluated at fixed parameters and . The polymer lengths are detailed in the respective legends.
It is important to note that the physical spatial extent of the polymer in the -direction is given by the transverse mean span (caliper size) of the chain, [43]. We estimate here the ratio for 2D SAWs on the square lattice [Figure S8, SM], which lies intermediate between the exact value for ideal Gaussian chains, [52], and that for self-avoiding rings, [53]. Consequently, the condition translates directly to , demonstrating that the chain must be strictly confined within a single terrace for the TA-to-SA criticality to manifest.
Figure S8. Extrapolation of the ratio between the mean transverse span and the radius of gyration, , plotted as a function of for SAWs on a 2D square lattice.
Conclusions. Our study reveals three possible thermodynamic scenarios for polymer adsorption on vicinal surfaces, depending on the terrace width: a single desorbed-to-adsorbed transition for small , analogous to the adsorption transition on flat 2D substrates; the same single transition for intermediate , but with the adsorbed phase presenting a crossover from the TA to the SA phase for large and increasing ; and the phase diagrams displayed in Figure 4 for large , where the three phases are separated by continuous transition lines. The presence of kinks in the steps does not alter this behavior for the parallel steps analyzed here.
Substantially, regardless of whether it experiences a crossover or a true phase transition, the polymer in scenarios and will ultimately end up in the SA phase at low , provided it interacts attractively with the step. If , starting from the desorbed phase and decreasing causes the chain to first adsorb onto the terrace and then attach to the step. For other interactions, a direct transition from the desorbed to the SA phase is possible. The desorbed-to-SA transition, alongside the multicritical point discovered in scenario , are novel features that invite future theoretical works to explain their underlying critical properties (particularly the large critical exponents observed here). Moreover, this highlights the importance of further investigating polymer adsorption on other experimentally relevant vicinal surfaces — such as those with non-parallel steps, rough steps, AB terraces, or alternating terrace-step bunches — as well as on complex crystalline surfaces featuring islands, mounds, or screw dislocations. The interplay between polymer size and terrace width revealed here indicates that polymerization at vicinal surfaces is also an interesting topic for future studies.
From a practical perspective, precise control of polymer conformation is crucial for tuning its electronic, mechanical, and biological properties for specific applications. Therefore, the existence of a stable SA phase, as unveiled here, indicates that such control can be achieved via attachment-detachment on engineered stepped surfaces (e.g., using lithography techniques). Furthermore, vicinal surfaces have long been utilized as templates for the self-assembled growth of nanowires formed by atomic chains decorating the steps [54]. Our results demonstrate that robust polymer nanowires can be produced at vicinal surfaces by setting the thermodynamic parameters deep inside the SA phase. Importantly, this approach can overcome the persistent difficulty in obtaining long wires caused by atomic defects (inherent to self-assembly), which frequently fragment the chains — since the wires will be formed via adsorption of long, continuous polymers.
Acknowledgments
The authors acknowledge financial support from CNPq and FAPEMIG (Brazilian agencies). This research utilized Queen Mary’s Apocrita HPC facility, supported by QMUL Research-IT [55], on which all simulations have been performed.
References
Supplementary Material for ``Polymer adsorption on vicinal surfaces''
Nathann T. Rodrigues; Tiago J. Oliveira
Introduction
This Supplementary Material (SM) provides additional details regarding the simulations of polymer chains adsorbing onto vicinal surfaces. We begin by defining the key quantities analyzed in this study. Subsequently, we discuss further data for polymer adsorption on surfaces with straight steps and wide terraces, followed by a complementary analysis of how kink sites and terrace width affect the polymer phase behavior.
Quantities of interest
To construct the phase diagrams for polymer adsorption, we first examine the fluctuation maps obtained from flatPERM simulations for . These maps correspond to density plots of the largest eigenvalue of the covariance matrix:
where and are, respectively, the numbers of monomer-terrace and monomer-step contacts. Regions of high fluctuation (indicated by brighter colors in the maps below) typically mark the phase boundaries. The average fractions of monomer contacts with terraces, , and with steps, , are important quantities for characterizing adsorption transitions. They are defined as:
where . In the commonly studied scenario of three-dimensional (3D) polymers adsorbing onto a perfectly flat 2D substrate, one has and in the ground states of the desorbed and adsorbed phases, respectively. A similar behavior is observed for when 2D polymers adsorb onto a 1D boundary. Therefore, in such simple geometries, these fractions serve as the order parameters for the adsorption transitions.
In the vicinity of the adsorption point (e.g., ), they are expected to scale as:
where and or . The scaling function assumes a constant value at , meaning that the curves of the effective exponents versus become independent of (i.e., they cross) exactly at the critical point when [1]. Note that, depending on the dimensionality of the transition (3D to 2D, 2D to 1D, etc.), only one of the fractions ( or ) will exhibit these scaling properties.
Other key observables are the relative fluctuations in the number of terrace contacts, , and step contacts, , which are defined as:
where or . At least one of these fluctuations will display a maximum near the critical point, the peak value of which is expected to scale as:
Moreover, the location of this maximum provides an estimate for the pseudocritical parameters or .
Results for vicinal surfaces with wide terraces and straight steps
As discussed in the main text, the phase diagram for polymers adsorbing onto surfaces where the steps are straight (i.e., free of kinks) and the terraces are sufficiently wide (ensuring the polymer interacts with only a single step) exhibits three distinct phases. These are: the desorbed phase (at small weights and ); the terrace-adsorbed phase (at small and large ); and the step-attached phase (at large and ). In the following subsections, we present a detailed analysis of the transition lines separating these phases and their corresponding critical properties.
Desorbed to Terrace-Adsorbed Transition
At low , increasing drives the polymer from the desorbed (3D coil) phase to the terrace-adsorbed phase. For perfectly flat substrates, this corresponds to the well-documented 3D to 2D continuous adsorption transition, which occurs at [2], with critical exponents [1, 2].
Figure S1 displays the effective exponents and as functions of , for . These effective exponents were calculated as the local slopes of the versus curves, where . The exponent approaches zero for small , as expected for the desorbed phase, and increases towards 1 (indicating polymer adsorption) for large . Conversely, consistently retains low values, demonstrating that terrace contacts overwhelmingly dominate over step contacts, which remain minimal across both phases. The crossing of the curves for different chain lengths confirms the presence of a continuous adsorption transition at . Furthermore, the value of at the crossing point aligns well with . An identical behavior is observed in the curves for other values of , giving rise to a desorbed to terrace-adsorbed transition line situated at , a finding also corroborated by the Flory exponents discussed in the main text. Along this line, we estimate critical exponents that fluctuate within the interval for SAWs up to 10240 steps, exhibiting no systematic trend to increase or decrease with varying .
The relative fluctuations versus , for , are plotted in Figure S1; analogous results are found
for other values of . The progressive build-up of these fluctuations with increasing provides additional evidence for a continuous transition. Consistent with the other observables, the locations of the maxima systematically converge to as . Moreover, scales as [see the inset of Figure S1], yielding crossover exponents in the range . This strongly suggests that the relation holds along the entirety of the desorbed to terrace-adsorbed transition line. Crucially, these exponents remain very close to the established flat-surface values of [1, 2].
Desorbed to Step-Attached Transition
For , increasing drives the system through a transition from the desorbed phase to the step-attached phase. An equivalent transition is observed when is held constant while increasing . Figure S2 provides a representative plot of the fluctuations versus , for . Notably, these fluctuations intensify with at a much faster rate than those associated with the desorbed to terrace-adsorbed transition shown in Figure S1. While such a rapid increase is typically a hallmark of first-order transitions, the corresponding probability distributions for the number of monomer-step contacts lack any trace of bimodality near the transition point [see Figure S2]. The monomer-terrace contact distributions [not shown] similarly exhibit no bimodal behavior. Thus, the desorbed to step-attached transition appears to be genuinely continuous.
The rapid growth in the fluctuations points to unexpectedly large critical exponents. Because every step contact inherently acts as a terrace contact, the effective exponents and are almost identical. Indeed, the intersection points of these exponent curves [analogous to those in Figure S1, though not shown here] provide nearly equivalent estimates for their pseudocritical values. These estimates, obtained for various as well as for fixed , are summarized in Figure S2. Extrapolating these values to the thermodynamic limit () yields . In cases where the data did not demonstrate monotonic convergence, only the largest chain lengths were utilized for the extrapolation. The critical points derived from these curve intersections are also in strong agreement with the transition points estimated independently via the Flory exponents and the maxima. The resulting desorbed to step-attached transition boundary is mapped out in Figure 2(a) of the main text.
As anticipated, the crossover exponents are similarly large. Figure S2 demonstrates that the finite-size values and occupy the same range as and . Upon extrapolating this data, we estimate the thermodynamic crossover exponents to be . This evidence strongly implies that the scaling relation is maintained for both and along this newly identified transition line.
Terrace-Adsorbed to Step-Attached Transition
For , the system undergoes a transition from a terrace-adsorbed state to a step-attached state as increases. In the extreme limit of , the polymer lies entirely flat on the surface, reducing the scenario to the ‘classic’ 2D to 1D continuous adsorption transition, known to occur at with exact critical exponents [3].
We observe a highly analogous scenario here for ; however, the behavior becomes more intricate at smaller values of . For instance, at , Figure S3 shows that across all values of , suggesting the vast majority of monomers are already adsorbed to the terrace. Yet, at the crossing point, , which is significantly smaller than the theoretical . As is lowered further, the values of at the crossings continue to drop, and these intersection points disappear entirely as approaches the multicritical point, as evidenced in Figure S3 for . Although a cursory glance might suggest the absence of a terrace-adsorbed to step-attached transition in this regime, we argue that this disappearance is purely an artifact of severe finite-size effects.
Figure S3. Results for vicinal surfaces with broad terraces and straight steps. Effective exponents (dashed lines) and (solid lines) plotted as functions of , evaluated at [(a) and (c)] and [(b) and (d)], for the chain lengths listed in the legend of (a). The exponents in the top panels were calculated using the standard scaling relation in Eq. (4), whereas the bottom panels utilized the modified scaling in Eq. (8).
Specifically, the average fraction of desorbed monomers (defined as ) is approximately 23% at the transition point for , but surges to roughly 80% for . Crucially, the scaling relation in Eq. (4) is physically expected to hold for only when the polymer is rigidly constrained to a 2D space, rendering the true fraction of step contacts per adsorbed monomer. In our actual 3D system, the properly normalized equivalent fraction is:
where represents the average number of surface-adsorbed monomers. By analogy with Eq. (4), it naturally follows that:
When the effective exponents are recalculated using this metric and plotted against , the expected curve crossings unequivocally reappear for all values of . Furthermore, these intersections occur at values of that match the critical points extracted independently from the Flory exponents. Even more compellingly, the values of at these new crossings correctly converge toward , as demonstrated in Figure S3 and Figure S3. The finite-size distortions caused by the desorbed monomers also heavily impact the normalized fluctuations. For example, if we attempt to estimate the crossover exponent from the maxima of the uncorrected curves [defined in Eq. (5)] versus , we extract unphysically small values near the multicritical point (e.g., for ). Additionally, the critical points derived from these naive maxima fail to align with those established by other observables.
However, by considering the fraction of step contacts strictly relative to the total number of adsorbed monomers, as mandated by Eq. (7), the corresponding fluctuation metric is properly corrected to:
Consequently, it is reasonable to expect that the maximum of this adjusted function scales as:
As anticipated, the crossover exponents derived via this corrected scaling are far closer to the expected . Nevertheless, in the immediate vicinity of the multicritical point, they still present slightly suppressed values — for instance, one obtains at — which implies that further, higher-order scaling corrections remain relevant in this specific regime. A substantially less plausible physical scenario would be that the exponent spontaneously deviates from solely in the neighborhood of the multicritical point.
Results for vicinal surfaces with wide terraces and kinked steps
Let us now consider vicinal surfaces where kinks are present along the steps, again assuming broad terraces. We have verified that the results for smoothly interpolate between those discussed above for straight steps () and those for the limiting case of . Therefore, we will concentrate exclusively on this latter case here, bearing in mind that analogous behaviors are observed for larger values of .
Figure S4 shows the Flory exponents at fixed (or ) for . As defined in the main text, our coordinate system is established such that the - and -axes are always perpendicular to the macroscopic step direction, while the -axis remains parallel to it. Consequently, the scaling behavior closely mirrors the straight-step scenario [Figure S4–(Figure S4)]. Specifically, in the step-attached phase, the chain becomes highly extended along the step () while remaining strictly confined in the transverse direction (). Furthermore, the surface exponents [Figure S4] and the crossover exponents (not shown) display the exact same scaling as the straight-step case, confirming that the underlying thermodynamics are functionally identical for both step geometries.
Figure S4. Results for vicinal surfaces with broad terraces and kinked steps featuring a distance between kinks of . Panels (a)–(c) present effective Flory exponents versus () for the indicated values of () kept fixed and a chain length of . Panel (d) displays the effective exponents (dashed lines) and (solid lines) versus , for across several lengths .
Results for vicinal surfaces with short terraces and straight steps
While the wide-terrace situations analyzed in previous sections ensure that the polymer interacts with only a single step, in practice, a polymer may possess an average radius of gyration large enough to simultaneously visit multiple steps. Hence, it is important to further investigate the thermodynamic properties of adsorption on substrates with very short () and intermediate () terrace widths.
The limiting case
As discussed in the main text, steps and terraces become structurally equivalent when ; thus, the step-attached and terrace-adsorbed phases becomes a single adsorbed phase. Since each surface monomer carries an identical weight , the desorbed-to-adsorbed transition strictly occurs at , where . This symmetric desorbed-to-adsorbed boundary is clearly corroborated by the fluctuation map displayed in Figure S5.
Figure S5. Results for vicinal surfaces with straight steps and a terrace width of . (a) Fluctuation map in the – plane for . Panels (b), (c), and (d) respectively display the effective Flory exponents (for ), the fluctuations , and the effective exponents (with ) as a function of , evaluated along the symmetric line . No appreciable difference is observed between and in panel (d). The lengths in (c) match those indicated by the legend in (d).
Figure S5 illustrates the variation of the Flory exponents with respect to , evaluated along . Notably, instead of exhibiting in the adsorbed phase, one observes across both phases. This apparent anomaly arises because our coordinate system maintains the -direction perpendicular to the individual microscopic terrace, while the global geometry of the vicinal surface constitutes a sloped staircase. Consequently, the local -direction is no longer perpendicular to the macroscopic surface plane, rendering it functionally equivalent to the -direction. Despite this geometric rotation, once the chain adsorbs onto the tilted plane, its spatial extension projects symmetrically across all relevant directions. Indeed, we find for small , and for large , exactly as expected for the desorbed and 2D-adsorbed phases, respectively. The crossing of the exponent curves precisely localizes the critical point at .
This finding is further corroborated by the behavior of the fluctuations [Fig. S5(c)] and the effective exponents and [Fig. S5(d)]. Therefore, since we are considering , the global critical weight is determined to be . Moreover, we obtain critical exponents and (for ) that are in excellent agreement with , as expected for 3D polymers adsorbing onto a flat 2D surface.
Intermediate widths
As indicated by the Flory exponents detailed in the main text, intermediate ’s fundamentally suppress a phase transition between the terrace-adsorbed and step-
attached phases, replacing it with a crossover. The fluctuation maps presented in Figure S6 for at increasing values of clearly illustrate that only mild fluctuations occur at the boundary between these phases (visible as the faint, horizontal bright stripe separating the large and small regions) when is small. Notably, this boundary feature disappears entirely for . Conversely, as increases, the boundary fluctuations intensify and the crossover region narrows. From onward — corresponding to widths larger than the average radius of gyration — no appreciable differences are discernible between the maps [see Figs. S6(d) and (e)]. This strongly implies that, for a fixed chain length , increasing the terrace width eventually forces the crossover to sharpen into a genuine phase transition. Alternatively, by holding fixed and increasing toward the thermodynamic limit, the system will asymptotically exhibit crossover behavior, even if small- data indicates the existence of a continuous transition.
Figure S7 presents a series of effective exponent measurements that confirm this geometric constraint. For relatively narrow widths (e.g., ) and chain lengths (yielding ), we observe deep inside both the terrace-adsorbed and step-attached regimes [Figure S7], verifying the absence of critical scaling as the system traverses the apparent phase boundary. Figure S7 plots versus for a wider terrace (); interestingly, the curves for (where ) display intersection points characteristic of a critical transition, clustered near . However, the curves for and (where the spatial constraint tightens to and , respectively) fail to cross the others, confirming that a crossover mechanism is recovered at sufficiently small ratios.
Conversely, if the length is systematically increased while holding the spatial ratio fixed at , for example, the hallmarks of criticality persist across all simulated sizes [Figure S7]. This provides robust evidence that a true continuous phase transition separates the terrace-adsorbed and step-attached states exclusively when is large, changing to a crossover when this ratio is small. Therefore, it is relevant to inquire: what threshold ratio demarcates these two distinct thermodynamic regimes? To answer this, we calculated the exponent deep inside the terrace-adsorbed phase for several , fixing while tuning . As shown in Figure S7, the effective exponents stop depending on the ratio and get close to zero only when . Thus, the terrace-adsorbed phase is genuinely devoid of a macroscopic number of monomer-step contacts only at these large ratios. When , the polymer’s extensive footprint makes it to traverse multiple steps, intrinsically imparting characteristics of the step-attached phase to the terrace-adsorbed state. Consequently, a sharp phase transition can only happen when . This constraint is consistent with the transition observed at [Figure S7] and the crossover seen at [Figure S7].
Since the concurrent crossing of multiple steps physically prevents the transition, we are naturally led to ask: how many steps must a polymer interact with before the critical behavior turns into a crossover? Equivalently, how many individual terraces does the chain span when ? While effectively gauges the root-mean-square spatial extent, the dimension that strictly governs step-crossing is the transverse mean span (caliper size) , which measures the absolute mean distance between the polymer’s geometrical extremities along the -direction. Although is known to share the same scaling as (i.e., vs. ), the universal amplitude ratio for 2D SAWs on the square lattice has, to our knowledge, never been quantified. To resolve this, we simulated SAWs on a square lattice for chain lengths up to , measuring both and , and their ratio is presented in Figure S8 as a function of . This specific choice of scaling variable effectively linearizes the data, enabling a extrapolation to the thermodynamic limit () which yields . This reveals that the geometric condition required to sustain a continuous phase transition between the two adsorbed phases is approximately . In physical terms, the polymer strictly cannot span more than a single terrace for the continuous transition to occur.
References
References
- [3]H. N. W. Lekkerkerker and R. Tuinier, Colloids and the Depletion Interaction (Springer Nature, 2011). doi: 10.1007/978-94-007-1223-2DOI
- [4]M. Ates, A review study of (bio)sensor systems based on conducting polymers. Materials Science and Engineering C 33 1853 (2013); doi:10.1016/j.msec.2013.01.035; E. Nakata, T. Nagase, S. Shinkai and I. Hamachi, Coupling a Natural Receptor Protein with an Artificial Receptor to Afford a Semisynthetic Fluorescent Biosensor. J. Am. Chem. Soc. 126, 490 (2004); doi:10.1021/ja035631i; S. Veerapandian, W. Jang, J. B. S. Et al. et al. Hydrogen-doped viscoplastic liquid metal microparticles for stretchable printed metal lines. Nat. Mater. 20, 533 (2021). doi: 10.1038/s41563-020-00863-7DOI
- [5]C. Yan, S. Hu, Q. Fei, B. Zhang and W. Wu, Advancing Drug Delivery: Design and Applications of MOF-Polyurethane Composites for Controlled Release Systems. ACS Omega 10, 43363 (2025); doi: 10.1021/acsomega.5c05108; E. A. Appel, M. W. Tibbitt, M. J. Webber, B. A. Mattix, O. Veiseh and Robert Langer. Self-assembled hydrogels utilizing polymer–nanoparticle interactions, Nat. Commun. 6, 6295 (2015). doi:10.1038/ncomms7295.DOI
- [6]Polymer Science and Innovative Applications: Materials, Techniques, and Future Developments, edited by M. A. Ali Al-Maadeed, D. Ponnamma, M. A. Carignano (Elsevier, 2020). doi:10.1016/C2018-0-00918-4
- [7]M. Bachmann, Thermodynamics and Statistical Mechanics of Molecular Systems (Cambridge University Press, UK, 2014). doi:10.1017/CBO9781139028516DOI
- [8]S. Pavlukhina, and S. Sukhishvili, Polymer Adsorption. In Encyclopedia of Polymer Science and Technology, 4th ed., edited by H. F. Mark (Wiley, New York, 2014).doi: 10.1002/0471440264.pst014.pub2
- [9]B. Kronberg, K. Holmberg, B. Lindman, Surface Chemistry of Surfactants and Polymers (Wiley, New York, 2014).doi:10.1002/9781118695968DOI
- [10]J. A. Plascak, P. H. L. Martins, and M. Bachmann, Solvent-dependent critical properties of polymer adsorption. Phys. Rev. E 95, 050501(R) (2017) doi: 10.1103/PhysRevE.95.050501; P. H. L. Martins, J. A. Plascak, and M. Bachmann, Adsorption of flexible polymer chains on a surface: Effects of different solvent conditions. J. Chem. Phys. 148, 204901 (2018). doi: 10.1063/1.5027270DOI
- [11]C. J. Bradly, A. L. Owczarek, and T. Prellberg, Universality of crossover scaling for the adsorption transition of lattice polymers. Phys. Rev. E 97, 022503 (2018). doi: 10.1103/PhysRevE.97.022503DOI
- [12]C. J. Bradly, A. L. Owczarek, and T. Prellberg, Adsorption of neighbor-avoiding walks on the simple cubic lattice. Phys. Rev. E 98, 062141 (2018). doi: 10.1103/PhysRevE.98.062141DOI
- [13]N. T. Rodrigues, T. Prellberg and A. L. Owczarek, Adsorption of interacting self-avoiding trails in two dimensions. Phys. Rev. E 100, 022121 (2019). doi: 10.1103/PhysRevE.100.022121DOI
- [14]N. T. Rodrigues, T. J. Oliveira, and T. Prellberg, Universality class of the special adsorption point of two-dimensional lattice polymers. Phys. Rev. E 108, 024146 (2023). doi:10.1103/PhysRevE.108.024146DOI
- [15]A. Jayaraman, C. K. Hall and J. Genzer, Designing Pattern-Recognition Surfaces for Selective Adsorption of Copolymer Sequences Using Lattice Monte Carlo Simulation. Phys. Rev. Lett. 94, 078103 (2005) doi: 10.1103/PhysRevLett.94.078103; Q.-H. Yang, H.-K. Qi, X. Yang, M.-B. Luo and J.-H. Huang, Simulation study on the adsorption of polymer chains on checkerboard-patterned surfaces. J. Polym. Sci. 60, 3107 (2022) doi: 10.1002/pol.20220321
- [16]M. Mödderl, W. Janke and M. Bachmann, Adsorption and Pattern Recognition of Polymers at Complex Surfaces with Attractive Stripelike Motifs. Phys. Rev. Lett. 112, 148303 (2014) doi:10.1103/PhysRevLett.112.148303; A. A. Polotsky and A. S. Ivanova, Effect of bending stiffness on the polymer adsorption onto a heterogeneous stripe-patterned surface. J. Phys. A: Math. Theor. 55, 375006 (2022) doi:10.1088/1751-8121/ac8589;DOI
- [17]M. Muthukumar, Pattern recognition by polyelectrolytes. J. Chem. Phys. 103, 4723 (1995) doi: 10.1063/1.470691.DOI
- [18]B. Pattanasiri, B. Liewehr and M. Bachmann, Surface pattern effects upon polymer adsorption. Physics Procedia 68, 105 (2015). doi:10.1016/j.phpro.2015.07.117DOI
- [19]H.-K. Qi, M.-B. Luo and J.-X. Chen, Simulation study of the selective adsorption and pattern recognition of polymer chains on patch-patterned surfaces. Colloids and Surfaces A: Physicochemical and Engineering Aspects 708, 136033 (2025). doi:10.1016/j.colsurfa.2024.136033DOI
- [20]A. J. Golumbfskie, V. S. Pande and A. K. Chakraborty, Simulation of biomimetic recognition between polymers and surfaces. PNAS 96, 11707 (1999) doi: 10.1073/pnas.96.21.11707. T. Bogner, A. Degenhard and F. Schmid, Molecular Recognition in a Lattice Model: An Enumeration Study. Phys. Rev. Lett. 93, 268108 (2004). doi:10.1103/PhysRevLett.93.268108DOI
- [21]See, e.g., Ref. [19] for a recent and more complete survey of literature.
- [22]D. Hone, H. Ji and P. A. Pincus, Polymer adsorption on rough surfaces. 1. Ideal long chain. Macromolecules 20, 2543 (1987) doi:10.1021/ma00176a038; R. C. Ball, M. Blunt and W. Barford, Can surface bound states be induced by interfacial roughness? J. Phys. A: Gem. Phys. 22, 2587 (1989) doi:10.1088/0305-4470/22/13/049; X. Xu, X. Sun and Y. Yang, Polymer adsorption between a rough surface and a flat surface for an ideal long chain. Acta Polym. 46, 145 (1995) doi:10.1002/actp.1995.010460208; W. Sung, J. Sung and S. Lee, Effect of surface undulation on polymer adsorption. Phys. Rev. E 71, 31805 (2005) doi: 10.1103/PhysRevE.71.031805.
- [23]H. Ji and D. Hone, Polymer adsorption on rough surfaces. 2. Good solvent conditions. Macromolecules 21, 2600 (1988) doi:10.1021/ma00186a049.
- [24]A. Baumgärtner and M. Muthukumar, Effects of surface roughness on adsorbed polymers. J. Chem. Phys. 94, 4062 (1991) doi:10.1063/1.460656.DOI
- [25]W. Gottstein, S. Kreitmeier, M. Wittkop, D. Goritz and F. Gotsis, Monte Carlo simulations of the adsorption of a single polymer chain on rough surfaces. Polymer 38, 1607 (1997) doi:10.1016/S0032-3861(96)00670-2. A. Striolo, Adsorption of model surfactantlike copolymers on nanopatterned surfaces. J. Chem. Phys. 125, 094709 (2006) doi:10.1063/1.2346676.DOI
- [26]T. M. Birshtein and O. V. Borisov, Theory of adsorption of polymer chains at spherical surfaces: 1. Partition function: diagram of states. Polymer 32, 916 (1991) doi:10.1016/0032-3861(91)90520-S; T. M. Birshtein and O. V. Borisov, Theory of adsorption of polymer chains at spherical surfaces: 2. Conformation of macromolecule in different regions of the diagram of states. Polymer 32, 923 (1991) doi:10.1016/0032-3861(91)90521-J; C. Y. Kong and M. Muthukumar, Monte Carlo study of adsorption of a polyelectrolyte onto charged surfaces. J. Chem. Phys. 109, 1522 (1998) doi:10.1063/1.476703
- [27]A. A. Gorbunov, E. B. Zhulina, A. M. Skvortsov, Theory of adsorption of macromolecules in cylindrical pores and at surfaces of cylindrical shape. Polymer 23, 1133 (1982) doi:10.1016/0032-3861(82)90367-6; A. Hanke, Adsorption transition of a self-avoiding polymer chain onto a rigid rod. J. Phys.: Condens. Matter 17, S1731 (2005) doi:10.1088/0953-8984/17/20/006DOI
- [28]A. Hanke, E. Eisenriegler and S. Dietrich, Polymer depletion effects near mesoscopic particles. Phys. Rev. E 59, 6853 (1999). doi:10.1103/PhysRevE.59.6853DOI
- [29]R. H. Levi, Y. Kantor and M. Kardar, Pinning and unbinding of ideal polymers from a wedge corner. Phys. Rev. E 96, 062132 (2017). doi:10.1103/PhysRevE.96.062132DOI
- [30]M. Blunt, W. Barford and R. C. Ball, Polymer adsorption and electron binding on rough and fractal surfaces. Macromolecules 22, 1458 (1989) doi:10.1021/ma00193a076; J. F. Douglas, How does surface roughness affect polymer-surface interactions? Macromolecules 22, 3707 (1989) doi:10.1021/ma00199a035; W. Nowicki, G. Nowicka, M. Dokowicz, A. Mńka, Conformational entropy of a polymer chain grafted to rough surfaces. J. Mol. Model. 19, 337 (2013) doi:10.1007/s00894-012-1546-5 T. A. Vilgis and G. Heinrich, Disorder-induced enhancement of polymer adsorption - a model for the rubber-polymer interaction in filled rubbers. Macromolecules 27, 7846 (1994) doi:10.1021/ma00104a047 A. Venkatakrishnan and V. K. Kuppa, Polymer adsorption on rough surfaces. Curr. Opin. Chem. Engin. 19, 170 (2018) doi:10.1016/j.coche.2018.03.001DOI
- [31]J. Krug and T. Michely, Islands, Mounds and Atoms (Cambridge University Press, Cambridge, 1995).
- [32]J. W. Evans, P. A. Thiel and M. C. Bartelt, Morphological evolution during epitaxial thin film growth. Formation of 2D islands and 3D mounds. Surf. Sci. Rep. 61, 1 (2006) doi:10.1016/j.surfrep.2005.08.004DOI
- [33]A. Saywell, J. Schwarz, S. Hecht, L. Grill, Polymerization on Stepped Surfaces: Alignment of Polymers and Identification of Catalytic Sites. Angew. Chem. Int. Ed. 51, 5096 (2012) doi:10.1002/anie.201200543; B. Cirera, Y. Q. Zhang, J. Bjork, S. Klyatskaya, Z. Chen, M. Ruben, J. V. Barth and F. Klappenberger, Synthesis of Extended Graphdiyne Wires by Vicinal Surface Templating. Nano Lett. 14, 1891 (2014) doi:10.1021/nl4046747 L. Grill and S. Hecht, Covalent on-surface polymerization. Nature Chem. 12, 115 (2020) doi:10.1038/s41557-019-0392-9DOI
- [34]R. Peköz, K. Johnston and D. Donadio, Adsorption of Dichlorobenzene on Au and Pt Stepped Surfaces Using van der Waals Density Functional Theory. J. Phys. Chem. C 116, 20409 (2012) doi:10.1021/jp306062s; J. Feng, J. M. Slocik, M. Sarikaya, R. R. Naik, B. L. Farmer and H. Heinz, Influence of the Shape of Nanostructured Metal Surfaces on Adsorption of Single Peptide Molecules in Aqueous Solution. Small 8, 1049 (2012) doi:10.1002/smll.201102066
- [35]Such as the Ehrlich-Schwoebel (ES) barrier [36] for crossing steps.
- [37]C. Kittel, Introduction to Solid State Physics, 8th ed. (Wiley, Hoboken, NJ, 2005).
- [38]F. Oesterhelt, D. Oesterhelt, M. Pfeiffer, A. Engel, H. E. Gaub, and D. J. Müller, Single molecule force spectroscopy by AFM indicates helical structure of poly(ethylene-glycol) in water. New J. Phys. 1, 6 (1999). doi:10.1088/1367-2630/1/1/006DOI
- [39]H. Dietz and M. Rief, Exploring the energy landscape of GFP by single-molecule mechanical experiments. Proc. Natl. Acad. Sci. U.S.A. 101, 16192 (2004). doi:10.1073/pnas.0404549101DOI
- [40]M. Rubinstein and R. H. Colby, Polymer Physics (Oxford University Press, New York, 2003).DOI
- [41]S. B. Smith, Y. Cui, and C. Bustamante, Overstretching β-DNA: The Elastic Response of Individual Double-Stranded and Single-Stranded DNA Molecules. Science 271, 795 (1996). doi:10.1126/science.271.5250.795DOI
- [42]C. Vanderzande, Lattice Models of Polymers (Cambridge University Press, Cambridge, 1998). doi:10.1017/CBO9780511563935DOI
- [43]E. J. Janse van Rensburg, Monte Carlo methods for the self-avoiding walk. J. Phys. A 42, 323001 (2009) doi:10.1088/1751-8113/42/32/323001.DOI
- [44]P. Grassberg, Pruned-enriched Rosenbluth method: Simulations of θ polymers of chain length up to 1000000. Phys. Rev. E. 56, 3682 (1997). doi: 10.1103/PhysRevE.56.3682
- [45]T. Prellberg and J. Krawczyk, Flat Histogram Version of the Pruned and Enriched Rosenbluth Method. Phys. Rev. Lett. 92, 120602 (2004). doi: 10.1103/physrevlett.92.120602DOI
- [46]P. J. Flory, Principles of Polymer Chemistry (Cornell University Press, NY, 1966)
- [47]S. Caracciolo, A. J. Guttmann, I. Jensen, A. Pelissetto, A. N. Rogers and A. D. Sokal, Correction-to-Scaling Exponents for Two-Dimensional Self-Avoiding Walks. J. Stat. Phys. 120, 1037 (2005) doi:10.1007/s10955-005-7004-3; I. Jensen, Enumeration of self-avoiding walks on the square lattice. J. Phys. A: Math. Gen. 37, 5503 (2004) doi:10.1088/0305-4470/37/21/002DOI
- [48]N. Clisby, Accurate Estimate of the Critical Exponent ν for Self-Avoiding Walks via a Fast Implementation of the Pivot Algorithm. Phys. Rev. Lett. 104, 055702 (2010). doi:10.1103/PhysRevLett.104.055702DOI
- [49]P. Grassberger, Simulations of grafted polymers in a good solvent. J. Phys. A: Math. Gen. 38, 323 (2005). doi:10.1088/0305-4470/38/2/003DOI
- [50]See the Supplementary Material appended to this Letter.
- [51]T. W. Burkhardt, E. Eisenriegler, and I. Guim, Conformal theory of energy correlations in the semi-infinite two-dimensional O(N) model. Nucl. Phys. B 316, 559 (1989). doi:10.1016/0550-3213(89)90058-8DOI
- [52]G. H. Weiss, Aspects and applications of the random walk (North-Holland, New York, 1994). doi:10.1137/1037108
- [53]V. Privman and J. Rudnick, Size of rings in two dimensions. J. Phys. A: Math. Gen. 18, L789 (1985). doi:10.1088/0305-4470/18/13/010
- [54]For recent references and a literature survey, see, e.g.: E. M. Smelova, K. M. Tsysar, D. Galkin, V. I. Zverev and R. R. Gimaev, Magnetic properties of oxide Mn nanowires on vicinal Au surface. J. Magn. Magn. Mat. 587, 171246 (2023) doi:10.1016/j.jmmm.2023.171246; W. Andrzejewska, P. Wojciechowski, M. V. Dobrotvorska, S. Mu- rawka, P. Sobieszczyk, M. Zelent and M. Lewandowski, Directional growth of iron oxide nanowires on a vic- inal copper surface. J. Phys.: Condens. Matter 36, 345004 (2024) doi:10.1088/1361-648X/ad3e58; S. Bel, Theoretical study of the Al and Pd nanowires on Cu(100) stepped surfaces: the role of atomistic mechanisms. J. Crystal Growth 673, 128388 (2026) doi: 10.1016/j.jcrysgro.2025.128388
- [55]http://doi.org/10.5281/zenodo.438045









