ASLT / Shelf-Life Prediction

1 Introduksi

1.1 Latar Belakang

Accelerated Shelf-Life Testing (ASLT) menjawab masalah: real-time shelf life test untuk produk pangan ambient (12+ bulan) tidak praktis untuk product development cycle 6 bulan. Solusi: jalankan storage at elevated temperatures (mis. 30°C, 40°C, 50°C) untuk accelerate degradation, lalu extrapolate ke ambient via Arrhenius equation (Mizrahi, Labuza, & Karel, 1970; Labuza, 1982; Hu, 2011).

ASLT framework: pilih quality indicator (browning, vitamin C, oxidation, sensory acceptability), monitor over time at multiple T, fit Arrhenius, ekstrapolasi ke storage T realistis. Kelemahan: bila mekanisme reaksi berubah dengan T (phase change, glass transition), Arrhenius tidak valid — perlu modified models.

1.2 Tujuan Modul

Menerima multi-temperature storage data + quality indicator vs time.
Fit zero-order, first-order, atau Weibull kinetics per temperature.
Fit Arrhenius equation untuk $k$ vs $T$.
Memprediksi shelf life pada storage T target ke quality end-point.
Audiens: R&D shelf life testing, regulatory compliance untuk best-before claim.

1.3 Posisi di Antara Alternatif

Pilih ASLT / Shelf-Life Prediction untuk multi-quality multi-temperature shelf life. Untuk Maillard browning specifically, pakai Maillard / Browning Kinetics. Untuk lipid oxidation, pakai Lipid Oxidation / Rancidity Kinetics. Untuk microbial shelf life, pakai Predictive Microbiology / Shelf-Life Modeling. Untuk survival analysis (failure time), pakai Weibull / Survival Shelf-Life Modeling.

2 Metode

2.1 Dasar Teoretis

Kinetic model classes:

Zero-order: $C(t) = C_0 + k_0 t$ (mis. browning $\Delta E$ early phase).

First-order: $C(t) = C_0 e^{-k_1 t}$ (mis. vitamin C, thiamine degradation).

Weibull/Hill: $C(t) = C_0 \exp[-(t/\alpha)^\beta]$ (mis. sensory perception nonlinear, Tedjo, Cornips, & Schiffer, 2002).

Arrhenius equation (Arrhenius, 1889):

$$\boxed{\, k(T) = A \cdot \exp\!\left(-\frac{E_a}{R T}\right) \,}$$

dengan $T$ dalam Kelvin, $R = 8.314$ J/(mol·K), $E_a$ activation energy (J/mol), $A$ pre-exponential factor.

Linearisasi: plot $\ln k$ vs $1/T$ → slope $-E_a/R$, intercept $\ln A$.

Tipikal $E_a$ (Labuza, 1982):

Reaction	$E_a$ (kJ/mol)	$Q_{10}$
Maillard browning	100–150	2–4
Lipid oxidation	40–100	2–3
Vitamin C degradation	50–80	1.5–2
Microbial growth	50–80	2–3
Enzymatic browning	30–50	1.5–2
Texture / staling	80–120	2–3

$Q_{10}$ — rate increase per 10 K (alternative ke Arrhenius):

$$Q_{10} = \exp\!\left(\frac{10 E_a}{R T_1 T_2}\right) \approx \frac{k(T+10)}{k(T)}$$

Shelf life prediction ke ambient $T_{\text{amb}}$:

$$t_{\text{shelf, amb}} = t_{\text{shelf, accelerated}} \cdot \exp\!\left[\frac{E_a}{R}\left(\frac{1}{T_{\text{amb}}} - \frac{1}{T_{\text{accel}}}\right)\right]$$

atau pakai Arrhenius $k(T_{\text{amb}})$ + integrate ke quality endpoint.

Acceleration factor (AF) = $t_{\text{shelf, amb}} / t_{\text{shelf, accelerated}}$. Tipikal 5–20× untuk $T$ shift 25 → 50 °C.

2.2 Persamaan Inti

Zero-order: $C = C_0 + k_0 t$
First-order: $C = C_0 e^{-k_1 t}$
Weibull: $C = C_0 \exp[-(t/\alpha)^\beta]$
Arrhenius: $k = A \exp(-E_a / RT)$
$Q_{10}$: $k(T+10)/k(T)$
Shelf life prediction: $t_{\text{amb}} = t_{\text{accel}} \cdot k_{\text{accel}} / k_{\text{amb}}$

2.3 Asumsi & Batas Validitas

Asumsi	Konsekuensi jika dilanggar	Cara cek di SQalytics
Same degradation mechanism semua T	Arrhenius linear fail	Cek $\ln k$ vs $1/T$ linearity
Suhu storage konstan	Drift kinetik	Catat T-logger
Quality endpoint well-defined	Shelf life ambiguous	Specify threshold
Multi-T cover wide range	Extrapolation bias	$\geq 3$ T levels
Replicate $\geq 3$ per timepoint	$k$ noisy	Modul flag
Tidak ada glass transition di range T	Mekanisme berubah	Limit T < $T_g$ + 20 K

3 Cara Kerja

3.1 Step-by-Step di SQalytics

Buka ASLT / Shelf-Life Prediction dari domain Mutu dan Analisis Lanjutan.
Muat tabel: Time, Quality_value, Temperature_C, opsional Replicate.
Pilih kinetic model: zero-order, first-order, Weibull.
Pilih temperature levels untuk Arrhenius fit (minimum 3).
Set quality endpoint (mis. browning $\Delta E$ = 5.0).
Set ambient T target prediction (mis. 25 °C).
Klik Run ASLT.
Tinjau: Tab Per-T Kinetics — $k$ per suhu + $R^2$; Tab Arrhenius Plot — $\ln k$ vs $1/T$; Tab Shelf Life Prediction — $t_{\text{shelf}}$ + CI; Tab Acceleration Factor — AF dari accel ke ambient.

3.2 Template Tabel Input + Contoh Data Sintetis

Vitamin C retention di juice powder (3 suhu, first-order):

Time_day	Temperature_C	Vit_C_mg_per_g
0	25	100.0
30	25	95.2
60	25	90.5
90	25	86.1
0	35	100.0
30	35	88.5
60	35	78.4
0	45	100.0
30	45	75.0

SYNTHETIC Data vitamin C juice powder — first-order degradation di 3 suhu akselerasi (25, 35, 45 °C). Endpoint: 80% retention. CSV setara: docs/assets/example-data/id/quality-advanced/template_quality_aslt.csv.

3.3 Contoh Luaran

Per-T kinetics fit (first-order):

Temp (°C)	$k$ (day⁻¹)	$t_{1/2}$ (day)	$R^2$
25	0.00164	423	0.998
35	0.00417	166	0.997
45	0.00958	72	0.994

Arrhenius fit ($\ln k$ vs $1/T$):

Parameter	Value
Slope $-E_a/R$	-8200 K
$E_a$	68.2 kJ/mol
Intercept $\ln A$	21.1
$A$	$1.5 \times 10^9$ /day
$R^2$ Arrhenius	0.999
$Q_{10}$ 25→35	2.54

Shelf life prediction (endpoint: 80% retention = 20 mg loss):

Storage T (°C)	$t_{\text{shelf}}$ (day)	(month)
4 (chilled)	~730	24
25 (ambient)	136	4.5
35 (warm)	53	1.8
45 (hot abuse)	23	0.77

Acceleration factor 25→45 °C = 136 / 23 = 5.9×.

ASLT Shelf-Life Prediction — figure 01 — Gambar 1. *Panel* (a) Arrhenius plot — $\ln k$ vs $1/T$ untuk 3 suhu (25, 35, 45 °C); *panel* (b) shelf life prediction curve vs storage temperature dengan CI band (vitamin C endpoint 80% retention).

Kesimpulan ringkas: "Vitamin C juice powder mengikuti first-order degradation dengan $E_a$ = 68.2 kJ/mol (within literature 50–80 kJ/mol vitamin C). $Q_{10}$ = 2.54 — rate 2.5× per 10 K. Predicted shelf life ke 80% retention: ambient 25 °C = 4.5 bulan, chilled 4 °C = 24 bulan. AF 25 → 45 °C = 5.9× → ASLT 23 hari pada 45 °C ekuivalen 4.5 bulan ambient. Recommendation: label best-before 6 bulan (1.3× margin terhadap predicted) dengan kemasan O₂ barrier. Lanjut ke Weibull / Survival Shelf-Life Modeling untuk failure-time analysis pada sensory acceptance endpoint, atau Compare Many Groups untuk uji statistik antar formulasi."

4 Kesimpulan

4.1 Relevansi Real-World

Best-before date calculation untuk regulator submission.
R&D screening new product shelf life cepat (6 minggu vs 24 bulan).
Packaging selection — uji O₂ barrier vs ambient.
Reformulasi guidance — extend shelf life dengan antioksidan.
Stability study farmasi — ICH Q1A R2 framework.

4.2 Where to Go from Here

Maillard / Browning Kinetics — focused Maillard.
Lipid Oxidation / Rancidity Kinetics — focused oxidation.
Predictive Microbiology / Shelf-Life Modeling — microbial side.
Weibull / Survival Shelf-Life Modeling — failure-time analysis.
IFT (2011) untuk shelf life guidelines.

⚙ Troubleshooting Cepat

Arrhenius plot non-linear. Mekanisme berubah dengan T; limit ke narrower T range.

AF terlalu tinggi (> 50×). Kinetics mungkin not Arrhenius; cek glass transition.

CI shelf life lebar. Add lebih banyak T levels atau replicate.

i Riwayat Revisi

Tanggal	Revisi	Penulis
2026-05-12	Draft v2 publikasi (KaTeX kinetics + Arrhenius + $Q_{10}$ + AF + APA Mizrahi-Labuza-Karel/IFT/van Boekel)	Claude
2026-05-12	Konversi MD → HTML (W5 quality-advanced batch)	Claude

4 Referensi

Mizrahi, S., Labuza, T. P., & Karel, M. (1970). Computer-aided predictions of extent of browning in dehydrated cabbage. Journal of Food Science, 35(6), 799–803. https://doi.org/10.1111/j.1365-2621.1970.tb01998.x
Labuza, T. P. (1982). Shelf-life dating of foods. Food & Nutrition Press.
Hu, F. (2011). Accelerated shelf-life testing. In Food and beverage stability and shelf life (pp. 415–432). Woodhead. https://doi.org/10.1533/9780857092540.2.415
Arrhenius, S. (1889). Über die Reaktionsgeschwindigkeit bei der Inversion von Rohrzucker durch Säuren. Zeitschrift für Physikalische Chemie, 4, 226–248.
van Boekel, M. A. J. S. (2008). Kinetic modeling of reactions in foods. CRC Press. https://doi.org/10.1201/9781420017410
IFT (Institute of Food Technologists). (2011). Shelf life of foods: Guidelines for determination. IFT Expert Panel.