Navigating the dynamic landscape of emerging industries demands robust analytical frameworks to decipher the complex trajectory of technological evolution. Traditional models often focus on external drivers or thematic linkages, overlooking the fundamental role of performance attributes, or “efficacy,” in steering technological progress. This article addresses this gap by proposing and demonstrating an integrated methodology that leverages the SAO (Subject-Action-Object) semantic structure and the Quality Function Deployment (QFD) model. The core objective is to construct an efficacy-driven technology evolution path, systematically revealing how specific performance goals catalyze and shape technological innovation. The lithium ion battery serves as our primary empirical case, offering a rich context to validate the proposed approach. By extracting and analyzing patent data, we aim to map the co-evolution of technical components and their associated performance characteristics, providing a nuanced understanding of the innovation pathways within this critical sector.
The study of technology evolution is foundational to innovation management, drawing from evolutionary economics to view technological change as a dynamic, path-dependent process. Research has diversified into examining various drivers, patterns, and methodologies, as summarized in the following table.
| Research Perspective | Core Focus | Typical Methods | Exemplar Insights |
|---|---|---|---|
| Driving Mechanisms | Policy influence, market demand, scientific breakthroughs. | Text mining, complex network analysis, S-curve analysis. | Policies and market needs create selection environments that shape R&D directions. |
| Patterns and Paths | Incremental vs. radical innovation, path dependency, knowledge diffusion. | Citation network analysis, main path analysis, knowledge graphs. | Technologies often evolve along defined trajectories, with knowledge flowing through core patent networks. |
| Trend Forecasting | Identifying future technological directions and potential disruptions. | Text mining with maturity metrics, fusion index analysis. | Convergence between different technological domains signals emerging innovation opportunities. |
While these approaches are valuable, a distinct gap exists in systematically linking the evolution of specific technologies to the refinement of concrete performance attributes. Methods like co-word analysis or citation networks reveal associations but often lack a structured framework to articulate how a “need for higher energy density” directly propelled the development of a specific cathode material for the lithium ion battery. This study posits that efficacy is a primary internal driver. To operationalize this insight, we integrate SAO semantic extraction, which efficiently parses technical documents, with the structured prioritization logic of QFD. This SAO-QFD fusion enables the automated identification of key technologies and the quantitative analysis of their relationships with evolving efficacy demands.
The proposed methodological framework consists of four sequential stages: data preparation, semantic extraction and clustering, key technology identification via QFD, and finally, the construction of the evolution path. The process begins with the retrieval and cleaning of patent data from authoritative databases like Derwent Innovations Index (DII). Relevant patents for the target industry, such as the lithium ion battery industry, are collected and preprocessed to remove noise. The textual data, particularly the “Novelty” and “Advantage” sections of patent abstracts, are then subjected to Natural Language Processing (NLP). The SAO structures are extracted from sentences; typically, technical subjects (S) and objects (O) from “Novelty” sections form the pool of technology terms $(t_i)$, while action-object (AO) pairs from “Advantage” sections yield efficacy phrases $(e_{ij})$.
To manage the high dimensionality of extracted terms, we employ a semantic clustering approach. First, the Word2Vec model, specifically the Skip-gram architecture, is used to learn vector representations of words by predicting context. The objective function is to maximize the average log probability:
$$ \max \frac{1}{T} \sum_{t=1}^{T} \sum_{-c \leq j \leq c, j \neq 0} \log p(w_{t+j} | w_t) $$
where $p(w_{t+j} | w_t)$ is defined using the softmax function:
$$ p(w_O | w_I) = \frac{\exp({v’_{w_O}}^\top v_{w_I})}{\sum_{w=1}^{W} \exp({v’_{w}}^\top v_{w_I})} $$
Here, $v_w$ and $v’_w$ are the input and output vector representations of word $w$, $w_I$ is the input word, $w_O$ is the output context word, and $W$ is the vocabulary size. The semantic similarity between two term vectors, crucial for clustering, is measured by cosine similarity:
$$ \text{similarity}(w_A, w_B) = \cos(\theta) = \frac{\mathbf{v}_{A} \cdot \mathbf{v}_{B}}{\|\mathbf{v}_{A}\|\|\mathbf{v}_{B}\|} $$
A hybrid hierarchical K-means clustering algorithm is then applied. K-means provides an initial, efficient partitioning of term vectors into $k$ macro-clusters by minimizing within-cluster variance:
$$ \underset{S}{\arg\min} \sum_{i=1}^{k} \sum_{\mathbf{x} \in S_i} \|\mathbf{x} – \boldsymbol{\mu}_i\|^2 $$
where $S_i$ are clusters and $\boldsymbol{\mu}_i$ their centroids. Subsequently, Agglomerative Hierarchical Clustering refines each macro-cluster. The distance between two clusters $U$ and $V$ is computed using average linkage:
$$ d(U, V) = \frac{1}{|U| \cdot |V|} \sum_{\mathbf{x} \in U} \sum_{\mathbf{y} \in V} \|\mathbf{x} – \mathbf{y}\| $$
This two-step process yields a hierarchical taxonomy for both technology and efficacy terms, organizing them into coherent themes and sub-themes.
The core of key technology identification rests on an adapted QFD model. The “Voice of the Customer” in traditional QFD is replaced by the clustered efficacy categories $E_q$. Their importance weight $(w_q)$ can be derived from frequency analysis or expert input. A technology-efficacy relationship matrix $(R)$ is constructed, where each cell $r_{pq}$ quantifies the strength of association between technology cluster $T_p$ and efficacy cluster $E_q$. This strength can be calculated as the weighted co-occurrence frequency across the patent corpus:
$$ r_{pq} = \sum_{i=1}^{N} \sum_{j=1}^{K} \omega_{ij} \cdot I(t_i \in T_p \land e_{ij} \in E_q) $$
where $\omega_{ij}$ is the local weight of efficacy term $e_{ij}$ in patent $i$, and $I$ is an indicator function. The absolute importance $(AI_p)$ and relative importance $(RI_p)$ of each technology cluster are then calculated as:
$$ AI_p = \sum_{q=1}^{Q} w_q \cdot r_{pq}, \quad RI_p = \frac{AI_p}{\sum_{p=1}^{P} AI_p} $$
Technologies with high $RI_p$ scores are identified as critical to fulfilling the dominant efficacy demands of the era.
Finally, the evolution path is constructed by analyzing the sequence of key technologies over consecutive time periods. The semantic similarity between technologies identified as key in period $\tau$ and those in period $\tau+1$ is computed using their vector representations. High similarity suggests a direct evolutionary link (e.g., refinement), while lower similarity may indicate a potential replacement or disruptive shift. Concurrently, the contribution share of each key technology to different efficacy categories is analyzed over time. This dual analysis—tracking technological succession and its shifting efficacy profile—reveals the efficacy-driven path: how a performance goal (e.g., improve safety) leads to Technology A, whose limitations in another goal (e.g., energy density) then drive the emergence of Technology B.
To demonstrate the practical application of the SAO-QFD framework, an empirical study was conducted on the lithium ion battery industry. A comprehensive dataset of 153,115 processed patent records was obtained from DII. From these, SAO extraction yielded foundational lists of technical and efficacy phrases. The subsequent hierarchical clustering organized efficacy terms into 12 primary categories and technology terms into 10 major groups, as summarized below.
| Efficacy Category | Representative Keywords |
|---|---|
| Energy Efficiency | High energy density, storage capacity, low self-discharge. |
| Charge/Discharge Performance | Fast charging, high efficiency, long cycle life. |
| Lifetime & Reliability | Long cycle life, low decay, stability. |
| Safety Performance | Explosion-proof, thermal runaway prevention, short-circuit prevention. |
| Cost Effectiveness | Low production cost, high cost-performance ratio. |
| Temperature Adaptability | High/low temperature operation, thermal management. |
| Environmental Friendliness | Easy recycling, low pollution, green materials. |
| Battery Management | State monitoring, charge balancing, health prediction. |
| Technology Category | Representative Keywords |
|---|---|
| Battery Electrode | Cathode material (LCO, LFP, NMC), anode material (graphite, silicon-carbon), electrode structure. |
| Electrolyte | Liquid electrolyte, solid-state electrolyte, ionic liquid. |
| Separator | Polyolefin membrane, ceramic-coated separator. |
| Battery Management System (BMS) | State estimation, cell balancing, thermal management, fault diagnosis. |
| Manufacturing & Assembly | Coating technology, stacking/winding, assembly process. |
A chronological analysis of efficacy importance weights revealed distinct phases in the development of the lithium ion battery. The early phase (pre-2000) was dominated by concerns over basic safety, manufacturability, and charge/discharge performance for consumer electronics. The 2000s saw a surge in the importance of energy efficiency and cycle life. The 2010s marked the era of electric vehicles, where energy density, lifetime, and advanced thermal management (temperature adaptability) became paramount. In the current era (2020s), focus has expanded to include cost-effectiveness, sustainability (environmental friendliness), and sophisticated battery management alongside the continuous push for higher performance.
Applying the QFD-based analysis to each time period allowed for the identification of key technologies driving these efficacy shifts. The evolution path for core components like electrode materials and Battery Management Systems (BMS) becomes clearly articulated. For the lithium ion battery cathode, the path demonstrates a clear efficacy-driven logic:
- Early Stage (Lithium Cobalt Oxide – LCO): Key Efficacy: High energy density. Provided sufficient energy for early electronics but was costly and had safety/cycle life concerns.
- Evolution to Lithium Iron Phosphate (LFP): Driving Efficacy: Safety, cost, cycle life. LFP addressed the safety and cost flaws of LCO, albeit with a trade-off in energy density, suitable for applications prioritizing safety and longevity.
- Evolution to Ternary (NMC/NCA) & High-Nickel Ternary: Driving Efficacy: Energy density, cycle life. To meet the stringent energy density demands of electric vehicles, ternary materials offered a superior balance. The subsequent shift to high-nickel formulations (e.g., NMC 811) was further driven by the need to maximize energy density and reduce cobalt content for cost and supply chain reasons.
Similarly, the anode technology for the lithium ion battery evolved from graphite (balanced performance) towards silicon-based composites, a transition primarily driven by the intense demand for higher energy density (capacity), though this introduced challenges in volume expansion and cycle life that defined subsequent R&D. The BMS technology path evolved from basic voltage/temperature monitoring (driven by safety) to encompass state-of-charge estimation, cell balancing, and prognostic health management, driven by the complex efficacy requirements of large-scale battery packs for EVs and grid storage, including reliability, longevity, and safety.
The constructed evolution paths substantiate the central thesis that efficacy is a powerful, intrinsic driver of technological change in emerging industries like the lithium ion battery sector. The SAO-QFD model successfully operationalized this concept, moving beyond mere thematic trends to reveal the “how” and “why” behind specific technological succession. For instance, it clearly shows that the transition from LFP to NMC cathodes was not random but a direct response to the escalating priority of “energy density” in the industry’s efficacy profile. The methodology’s strength lies in its ability to decompose the monolithic concept of “performance” into specific, measurable attributes and quantitatively link them to clusters of patented technologies over time. This provides a more granular and actionable understanding for R&D strategists, highlighting not only which technologies are prominent but also which performance gaps they are attempting to fill and what trade-offs are being made.
However, the study acknowledges certain limitations. The primary limitation is the lack of a quantified mapping between the scale of a technological advance and the measurable improvement in a specific efficacy metric. For example, the analysis can indicate that high-nickel cathodes are linked to “higher energy density,” but it does not quantify the potential increase from, say, 250 Wh/kg to over 300 Wh/kg. This connection, while logically sound, remains qualitative. Future research could integrate numerical performance data from scientific literature or technical reports to enrich the QFD relationship matrix with intensity scores, moving from association to predictive correlation. Furthermore, the model currently analyzes historical data; incorporating real-time data streams and predictive analytics could enhance its utility for forecasting future inflection points in the technology evolution of the lithium ion battery and other sectors.
In conclusion, this article presents a novel, efficacy-driven framework for constructing technology evolution paths in emerging industries. By synergistically combining SAO semantic extraction with the structured prioritization mechanism of QFD, the method systematically identifies key technologies and elucidates the performance attributes propelling their development and replacement. The empirical application to the lithium ion battery industry validates the framework’s effectiveness, tracing clear evolutionary pathways for critical components like electrode materials and battery management systems. These paths illustrate a consistent narrative where technological iterations are purpose-driven responses to shifting priorities among competing efficacy demands such as energy density, safety, cost, and lifespan. This approach offers academic researchers a refined lens for studying innovation dynamics and provides industry practitioners with a structured methodology for strategic R&D planning, patent portfolio analysis, and anticipating future technological trends. Ultimately, understanding the efficacy-driven path is key to navigating the complex innovation landscape of transformative industries like the lithium ion battery sector.